OCR of the Document

View the Document >>

Peter T. Kirstein, University College London ARPANET Project Annual Report, February 1977. Unclassified.

A D—AU’I fl9 UNCLASSIFIED - UNIVERSITY COtS LONDON ENGLAND DEPT OF STATISTICS —— ETC r e 17 2 UNIVERSITY COLLEGE LONDON ARPANET PROJECT U FEB 77 P T KIRSTEIN NUOO1 l4— 7le—C Q28o TR—S NL r r uru p ‘ N - _ _ _ I _ L J S A A A A 4 25 3l 1 1 # i I 2 O j u m i _J 1 1 r L d n t new p r o j e c t t 1 - i t PS started in 1975 our prac ti r 1 l r t v A lthough Project 5 l 2 22 c p o n 3i tn programs began only in 1976 Durir L L r vern i i th i s r i o a r’ T R EPO RT DOCUMENTA T ION PAC E I 4 Nl I Q R T N U M L R — U n i v e r si ty Coil ege London AHP tNET Proj ect - - - Profcssor Peter T Kirstein 9 — S TY PE or REPORT E P E R F OR M I NG ORG CONTRACT OR ID Prof Peter T Kir s t e t epartflUflt of Statistics and Co puter Science University College London WC1E 6BT UK 14 C O N T R O L L I NG O F F I C E N A M E A N D A D D R E S S REPORT GRANT DEC 76 NU M ER N UMB L F I -- P E R F O R M I N G O R G A N I Z A 1 ION N A M E A N D A D D R E S S f en ce Advanced Research Proj ects Agency 1400 Wilson Boulevard Arlington Vi r g i n i a 22209 - 6 PEHIOL C O V E R E D 1 JAN I -31 AN U L P R O G RA M E L E M E N T P R O J E C T A R I A A WO R K uNIT NUM BE RS 12 REPORT differen t from ControZIIn Of fice M O N I T O R I N G A G E N C Y N A M E A A O D RES S I 1 DATE 1 3 J M B E R O F 187 15 TASK II - A GE S SECURITY CLASS of thl report UNCLASSI FlED ISa DECLASS IFICAT ION DOWNGNAOING 5CR E DU L E _____- TI IA - D I S T R I B U T I O N S T A T E M E N T of thla Report No 1 - Il DISTRIBUTION S T A T E M E N T nv d i n P b o r k 20 1 hat of t h Il renl Irorn Report E J 1f ThE 22 t9fl SEP IA SUPPLEMENTARY T EFU L L A NOTES 19 KEY QR DS C oo l i no e on r r ’r n r r fC If n e r e ’ acr loeoSIIy by b h ‘rd r r i r r Ir o - rnet working S%cITCI I Fi le Transfer Facs imile Network Measurenients ork Usage ‘on it o r i n g ItT EPSS Packet S a te l l i te Experiments N etv 20 A BS I RA E I hr h I o r b If r e r r a r y nd I d e n t I f y Thi s 1’eIx rt i ’ a - I tna l fl fl’t On a CL f l t I 1 1 1 i’ 11 t h e 01 f i c e of Naval Hf - rc i N000l 4 --7 1— C— 0280 wh i ch tenni n u t ed on 30 Sf ‘pt u1I lvr 1976 and a the ii rst quar t eç lv report on Toni rae I N000l 4— —G—-0005 ich began on 1 itober 1976 Research a ’t i v i t i e s coVered i n th is k I Urt ’fl t i n c l u d e the I O f l h I ct 1 1t Of I HI S and n etwo r ks t h e h ye l Opflk ‘UI Ed a syst SW I TCH Ic 11rt ua n ’t Ork connec t ion and our connect i I I I 11 t ilL Ilk l st hU I p1 S covered in particu l ar Oh c Exper imental Packet Swi tchl ’ l Se r V I ’f ’ T ‘ 1 DD 1473 E D I T I O N OF I NOV 65 IS OI1SOL t I E SECURITY 7 _ _ _ _ _ _ _ CL A S SI F IC A l ION OF T ill S P A G E llb D a t a F I - - -- -- SECU RITY CLA SSIF ICA 1 I N OF T H I S PA E NJ 1 1 _ _ _ _ _ _ _ _ _ _ _ _ __ __ _ _ _ _ _ _ _ _ _ _ _ _ _ ar e the 1 ket s a t e l l i t e project and the an iiss10n Control rogr other ARPA up or1 E d ‘J X P ex 1x ’riir ’ntS Th ese haVe particular rel evance to our nE su rec iit and F i n a l l y projectS and so ax e treated in some depth through the AIWANET of and UK use are described facsimile a c t i v i ti e s dra be may that conclusions ed A list of evaluat i from our WL TIP is the of report research activities during 1976 is given at the ehd det ail 1T3 LMNIm f’ r Wh e Section SHS ction i j SECURITY — _L - -- - — o T C A S S I F I C A T ION OF IRIS PAGE Il en Da t a F ‘ 1 ‘tn ‘ BS’ ’ I ACT T h i s r e p o r t serve s as a f i n a l r e p o r t on a c n- r -t from the Office of Naval Research NOU 0l4—74—C --O 3fl - ‘- h i c h t e r m i n a t e d on 33 S e p t e m b e r 1976 a na quarterly s th E report on Contract N O U 0 1 4_ 7 7 _ f l - O 1 - l ’ began on 1 October 1 Y 6 Resecxrcri a c t i v i t i e s t h i s d ‘usc’ st i nc l u de t h e i n tc r c o nn e - ’t i o n n e L o r k t h e d eve l opm ent ef a sy st -- n e t w o r k c o n n e c t ion and our conne ’ti n to of he first r ” -r n o st s ‘H O f f i c e E x p e r i m e n t a l P a c k e t wi t ch ed S e r v i E e cove rt d i n p a i t i c u l i r de t a i l are t h e P i ck e t c nd u F ° Th’ - t I L - T’ tc P r o j e c t ‘xnd the Transmission Control Pr sr i’ expe r iments These i ave part ic u l a r r e l e v a n c e t c a er ARP A s iu p e rt e d p r o j e c t s and so a r e t r e a t ci in t Fi nd v o u r mea u rc en t and f a c sir i Ic ac t i v t 1’ des c r i b e d and UK use of A P A P T t h r e u h t h e T ’ ’ I e va l u a t e d A l i s t of c o n cl u s i ns t h a t r c i V be d ii r o our reSe3rCh activit ies during 1976 is c iv en at t Sr I ef Lhe repor t - - — i PA GE TA BLE OF CONTENTS TITLE PAGE ABSTRACT LIST OF FIGURES LIST OF ABBRE V IATIONS USED IN TUE REP ORT INTRODUCTION 1 1 1 2 1 3 2 Purpose of the Report Aims of the Project and their Relation to this Report 2 Institutional Arrangements 4 Introduction 5 Operating System Measurement and Improvement 8 2 2 The UCL Configuration 2 4 The Attachment of the RL IBM 360 195 2 3 2 5 The Attachment of the Royal Signals and Radar 2 6 The Attachruent of the ULCC CDC 6000 7000 Complex 2 7 5 9 Establishment GEC 4080 10 The Attachment of the Cuiham ICL System 4 72 13 11 SWITCH - AS A NETWORK It ITERFACE 3 1 3 2 Introduction 15 3 2 1 Message of Character Termina l Service 1 3 2 3 Flow Control Interactive Terminal Service 3 2 2 3 3 4 — UCL SYSTEMS DEVELOPMENT 2 1 3 1 - 1 16 A Single Service Level 17 18 F i le Tr ansfer Serv ic e 3 3 1 6 18 The File Transfer Service in Detail 20 THE INTERCONNECTION OF COMPUTER NETWORKS 22 4 2 26 4 1 Survey and Back ground 22 UCL Work in Mapping Virtual Call Protoco l 4 2 1 Introduction 26 4 2 3 SWITCH File Transfer Mapping 32 4 2 2 Interactive SWITCH Mapp ing I tr 1r 27 c_ rnfl 11tff’ m - I r— t ‘ r - 4 3 4 4 5 UCL Work in Internetworking using Strong End—to—End Pro tocols 35 1977 78 37 UCL Approache s to Internetworking p l a n n e d f o r THE TRANSMISSION CONTROL PROGRAM 5 1 5 2 Introduction 5 2 1 5 2 2 5 3 5 2 3 5 5 6 45 Exp er imen tal Software 49 5 3 1 50 The UCL Implementation Internetwork Experiments with TCP 51 53 Sim u l a tion St u d ies Conclusions Introduction 6 2 UCL Hardware and Line Handlers 6 4 The Packet Hasp Experiment 6S File Transfers 59 6 3 6 5 Software for EPSS Attachment Future Support of Asynchronous Terminals and THE PACKET SATELLITE PROJECT 7 1 Objectives of Broadcast Satellite Techniques for 7 2 The PSP Configuration 7 4 The UCL Measurement Tools 7 3 7 5 Packet Switching 71 General Progress during 1976 — 72 76 An Overview 77 Traffic Generation and Data Acquisition 7 5 1 7 5 2 7 5 3 7 5 4 7 6 42 Configurations and Methods UCL ACTIVITIES WITH EPSS 6 1 7 42 A ims and Env ironment Experimental Results 5 3 2 5 4 41 Experiments and Tools 7 5 5 79 The Environment 79 The Traffic Generator 80 Data Acquisition and Reduction 82 Expanding the Message Building-in Response Control of the Traffic Generator and the Experiments - — - 83 7 6 1 7 6 2 7 7 7 8 8 7 6 3 8 2 Introduction M onitorinq of P S TN 88 Usag e 95 8 5 Network-Level Performance Measurement 99 Characteristics of a Specific Host 104 8 5 1 107 8 5 3 107 Introduction Tool Facilities M easurement Results 108 109 Line Measurements 111 EXPERIENCES W I T H UK USAGE OF ARPANET 9 1 9 2 Introduction 9 2 1 9 2 3 9 2 4 9 2 5 9 3 9 4 117 Ana l ysis of Usage during 1976 9 2 2 10 88 MEASUREMEN TS 8 5 2 9 84 The Application of the ieasurement Tools Planned f o r SATNET Measurement of Usage via PDP9A 8 6 84 Gateway—Level Functions 8 3 8 4 83 The State of the Measurement Tools NETWORK 8 1 The Function i n i S t r u c t u r e Controller—Specific Features 9 2 6 118 ARPANET Sites Used T eic of Network Use Problems encountered in the Use of ARPANET lodc of Access to the Network Connection Times 9 3 2 122 124 125 125 Summary 121 UK Collaborative Research Projects usinq ARPANET 9 3 1 120 hl i qh— Activity Projects Low—Activity Projects 127 133 Conclusions 13 FACSIMILE ACTIVITIES 10 1 10 2 10 3 Introduction 139 The System Overview 141 14 The User View of the System 10 3 1 10 3 2 10 3 3 Introduction The Facsimile Terminal 14 The User Dialoque for DocumentTran snission 148 --- - - - - - -—-— — - — - - —-- - — - --— - ----- - - - _ _ 10 3 4 10 4 10 3 5 10 4 1 10 4 3 -- -- 10 5 2 10 5 3 - The Dialoyue for Text and Facsimile R e t r ieval 151 Generation of Message Code Numbers 155 156 Characteristics of the Datacomputer 156 Di rectory Struc ture Privacy and File Access 157 Use of the Datacomputer 158 UCL Implementation 10 5 1 11 Storage and Archival at CCA 10 4 2 10 5 -—- — - — 160 The Hardware Confi guration 161 The FAXSYS Software 161 The Present Status 163 CONCLUSIONS 165 REFERENCES 171 APPENDIX A APPENDIX B APPENDIX C List of Low—Activity Research Groups 174 178 List of High-Ac tivity J esearch G ro up s Papers Presented during the Reporting 179 Period — -- --- - - — - --- -w IS UCL Configuration at the end of 1976 4 2 P a r t i a l C o n f i g u r a t i o n of N e t w o r k s and G a t e w a y s accessible from UCL by mid—1977 S o f t w r x r e m o d u l e s m a p p i n g b e t w e e n EPSS and ARPA NE T 2 l Experimental TCP Configurations 3 4f Transmission Queue Delays in the UCL TIP 52 5 5 Throughput for One and Two Connections 54 5 7 Conservative 5 1 5 2 Timestamp Packet Format 5 4 Round Trip Times for TCP Experiment 5 3 5 6 48 51 Throughput to Stanford BBN 5€ Scheme 60 A b i l i ty of V a r i o u s Schemes to meet D i f f e r e n t 5 8 7 1 Criteria 61 Proposed PSP Configuration for 1977 75 Current PSP Configuration 7 2 7 3 7 4 7 5 7 6J 74 Control Paths in the Satellite Net 85 90 Histograms of Gateway Measurements 91 8 1 Monitoring Performed as Function of Time of Day 8 3 Usage cf ARPANET by Various Research Groups 8 2 QUES Monitoring Performed 97 98 101 Number of Loq ins ncr Hour 8 5 92 100 Usage of V a r i o u s ARPANET Hosts 8 4 102 Number of Lornins nor Kour on PDP —A 105 I 1 s ee of RL 360 195 b ’ Various Groups V i a U C L P D p 9 106 Histogram of Types of Packet as a Function of Time 112 8 6 8 7 8 9 8 10 Main Log of Packet Descri ptions 113 10 1 Overview of an Idealised Facsimile System 143 10 3 Schematic of UCL Facsimile Terminal 10 5 b a nque for Accessing the Message Processor Dialogue for Retrieving Facsimile Files from the Datacomputer Messaqe RFNM Delay Time 8 11 10 2 145 147 Dialogue for Sendina a Document 10 6 _ 114 Schematic of UCL Facsimile System 10 4 _ RE S 2 1 4 1 _ ElI _ _ _ _ _ _ 152 153 154 _ _ _ _ _ _ _ _ _ _ _ _ ____________________________ 10 7 Directory Structure for Facsimile Tree Data-Language Commands to Retrieve and File Datacomputer Login Procedure 10 8 10 9 10 10 The UCL—ISI-CCA-ARPA Experimental Facsimile Configuration - - - - - - - 157 159 160 162 - -- - “ SOME ABB REVIAT I ONS —- - - USED IN THE REPORT BBN Bol t Beranek and Newman DCA Defence Commun i c ati ons Agency EPSS Exper imental P acke t Sw itched Serv ice IMP Interf ace Messa ge Processor MOD Mini s try o f Defence NLM Nat iorral Library of Medicine PSDN P u b l i c Switched Data Network PsP Packet Satellite Project PSTN P acke t Sw i tched Telephone N etwork PSU Packet Switching Unit Ru th e r f or d Lab o r a t o r y _ _ RSRE Royal Sig n a l s and Radar E s t a b l i s h m e n t SIMP S a t e l lite Imp TCP Transmission Contro l Program TPU Trans mi s si on P rotocol Uni t UCL Un ive rs ity C o l l e g e London UCLA University of Califo rnia at ULCC Uni vers ity of London Computer Centre _ _ _ _ - - - Los Angeles - - CHAP TER i Pl i CL dOSO T h e Urx i versitv in 1973 fur - or fl OT 1 _ _- -- INTRODUCTION or E ‘ cl i c c c’r on IAi 4 iana f c c -n t r ct f r o m f Pro3 PA ‘ c o r i c e of t• Navc 1 0001 4— 74—C—C i 8 0 w n r e n o w n a u to Sc u t e o n e r 30 1976 a new contract N ’J0i4—77—G —0 005 st rtcd on october 1 197€ This ronurt s e rv i S t h e r e f o r e a s a fLIal resort on the Research first contract ann re t at as rae fr rst c i u n r t e r i y on the second The project s also partially f r eu i’ the t’ritish Lr ra ry Pt dkA Yit ai- -d— 093 n o UK i in 1 s tr v or t u a _ e M i Ce rtract AT ’20 4 7 0 6 4 and r u e oX Sc enc Reseai ’ cn B RG 09811 and P f 17022 ann a i a nn u a a GRANT p ci’ t is t - L u l u r C u l S o an renort or r nc o n r r t ’ nun cc t a c a- ae tre previous renor ts 1 r s t t ir 97 nU Krrst eir l976A support The r’ p o s t d ficc has been aroci cd or t h e UK P r o j e c t ann tar s r eport a a s- a r c a the progress cf work T f C n r to’ It c P a L t S tcllite l n f oj T u t J ag t h em of i teilite E x c o ri m e n t C h a pt e r 7 and t h e TCP C ocrter 5 re rcated fn - ar tic ula r detail Th a is b ecause t o n e s u- j o c t a a r e or n rt lcui ar rc cr- ace to other subjects t a o i- ck -t ARPA s u r p o rt e a p r oj e c t s and hi a a pert is a final ON R N O Q Q 1 4 — 7 4 — C — 2 a 1 2 of report for t h e aroject o their Pcfation to thr s R e p o r t b u r ir n j L n e p r a v io u s y e ar s of ah project ‘c h a e i nv e s ti a t e d som e r c r i n c t les f o r i n t o r c o n n e L t iSC hosts and con o rer networks We had i n v es t i g at e d R u t h e r f o r d L a b o r a t o ry r i London Computer Ce tre l b 1 the- problems of connecting t h e 3 1u 19 5 th e U r d v e r s i t y of LOCT CPC 6000 7000 t h e R o y a l S ior i ni s and Radar Esta d i a u ’ont PSEE bcC - 080 and the Uulham ICL Sys tems 4 7 2 o nt A R r ANi i rrain p UUL ccn puters as the front— ends Wh ale a ll t e i nu LV t c u i l co a u ter s a n d been attached $ ep ar a t e y a ir i ng t h e p a s t y e a r we h a ve developed a s y s t e m f o r c o n n e c tin g a ll of t h e m s i n l u L t a r ’ d 5 I the same front-end co O u t e r c o m p l e t e d ny t r c end of discuss now the UCL l h s d e v ei c o en t t h is r ep o rt in a p e r i o d S J Stt S hardw ar e system software has r e q u i r en o n t o A RPAN i ’i na d I t a n ac-f l f l i tion of a t t a c h e d to was cc sen tinfi v In C h a p t e r 2 w e as r eg ards th e t t c a ent of n os ts The work i c-v e lo p e d u u s i n q u 1 9 76 new f o r m of v i r t u a l network c o n n e c t i o n as t h e n u n a e a of h o s t s to be a t t a c n c n grew t o f o u r The i m p le m e nt a t i o n of t h e r r w s y at em ca tad SWI’iC j in d i s c u s s e d i n C h a p t e r 3 ---- - - - - -- - - -- - - --- - --- - -- - - - - - -- -- -- - ---- — — - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ __ __ __ _ - T -ie logical nor’s st ep ‘ a s t o i n v e s t i g at e t h e pr oblems of nterconr ecting networks Here we had analysed during the previous the facilities roc uircd in gateways between networks and had started implementing a connection to another specific network p e r io d the UK Post Office P0 Experimental Packet Switched Service EPSS Broomficld l975 Tacre have been significant changes in the last year due to the acceptance of X25 i ntorface CCITT 1976 as an agreed standard for access to PTT networks whi ch can be u s ed w i t h aet’ ’arks This is a virtual call network - ora e modif ication for interconnet -tion of Our general approach is in fact a virtua l and so is compatible with this standard of services c ill mapping Our approach to t h e p r o v i s i o n throug h concatenated networks is developed further in Chapter 4 O ne ap p ro ach i the to comm un icatton between hosts or different networks implem entat Ion of software to a comm on We interface standard h tvc investigated a s ccific protocol TCP de siened for communicating between such hosts These ex p c r i n en”s n w h i c h have involved both Stanford University and Bolt Bornnc de t -a t i in Chapter 5 and Newman a r e d i s c u s s e d in some We have c o m p l e t e d quantitative experiments and fso simulated some n c d i f i c a t i o n s w h i c h h a v e been sueqested to I n pr e v e the p e r f o r m a n c e of sy s t o c t a u s i n g t h e p r o t o c o l have During t h e p a s t year we have bc-come increasingly involved with LPSS An exper ime’ ta interconnection between EPSS and ARP AN I T has been aevised w rtn one D P9 performing m appings between ARPANET a nd EPSS protocols The present UCL sy tem is not too advanced it perform s EPSS mappings at a very low level which does not use EPSS ef fi ienti y However t h in was the connection first adopted by RL and it has allowed us to get some experience with the problems of using EPSS We are now running experimental sessions for UCL users of the RL 360 195 usir ’ t re ARPANET EPSS and the tJCL gateway This development will h extended considerably during 1977 when mapping of the hig her leve l EPSS protocols will be achieved This work on EPSS is discussed in Chapter 6 w Although an important new project t h e A t l ’ L i Pacact ’ Sut f Project PSP started in 1975 our pract ical i nv iren#crt in li r e tao exper imental program s began only in 1976 Duriro this p r ioC a gateway computer was installed in University College London We began implement ing tools to gener at e t r a f f i c acr oss tb packet satell ite network to control the experiments and to collect ara Our act ivities in trn s flowing between the gateway computers area are discussed in Chapter 7 has been v ’rv - rtcnat ve In the previous annual report we discussed cu A key component of the UCL activities network measurements tools for access control and measuring terminal access c o n t ro i the public switched telephone network P5Th Dur in y throu gh the past year these tools have been developed further and fairly extensive and var ied measuremen ts have been made a nd anai y ed although P5Th monitoring has been restricted We y shortage of c o m u u t c r t t m c have complete records of all traffic using the UCL RL link i n either direction w c n e v er th e RL m a c h i n e has onen coanect e to AP PAEET The information is output on paper tape and has been k e p t s i n ce th e beg inning of the UCL ARPANET project although it n et p o ss ib l e to analyse it until 1976 Another form of measur ement that was taken during the year was that of line traffic passing aerace r this b aa been carried out in conjunction w ith the experiments on end— to-end protocols and has clarified our -caPe rstandarig of wh i cn performance pr ob lems depended on nest-nest interactions nra on the interface between host and data network operatiena FinaUy some m easurements have been made on the actual performance of PIf ferent ne two rk c o n t r o l programs themselves descr ibed in some detail ll the above projects are in Chapter 8 - The UCL node a’ s been almost unique on A R PA N E T in tha t it has a v e r y wi r e s p e c t r u m of u s e r s both i n t h e UK and t h e US It is p o s s i b le to t d c n t i f y t h e user c o m m u n i t y f a i r l y c l e a r l y because of the operational rules set up at the beginning of the prcUect see C h a p t e r 2 of K ir s t e i n 19 7 6 A u n d e r w h i c h any UK user m u s t o b t a i n e xp l i c i t a thurisation from the Governing Committee of the project S im i l a r l y every US user of tnt UCL node must obtain explicit pc rn issicn from ARPA As a result of these authorisation procedures it rs po ssible — - — — - - -- — — ‘ -- —- to f - tlrl l kn- w w a c u a c o r n te l v be p as e JectS d i f r e n s vo E a r ’ ’ A’Ae a c k ct cc-nc L u s i on s r n C a ’ sor - r o u p -- 1h ‘scd inha rrirv f sic i a irm ’s -cr n an of A list cd c an d P - - t from ati • c t v r i 0 rn s c ir ztrc tac- wt ti-r c -a ra c o n r ’ct o v er this r ot 1 a th ioU-c UCh a Lea c ‘sh e t Cr r i 1 i a - in f or tac e ’s’ e e f a a i 11 nd B n i aci e u- C a a p t er ‘ f l c k d u -” f- c s t n i l e i r u S a -iC - c n - t i n c -V term ti c- f a c e- l- 7 10 ‘R sod nai the t ct - -’t- r an d r i oQI t 1vc- 50CC re C h a p t - r r n - a r to li-N and t I m e N- - c ar c wai c -n u sL t rteP c ISt ’ U a u e n sri a r a r n o a min t sw ’sc-w i ne t wer-U m a are ye cii -i LL y ova l n i 1 l et - - 1 cancer Grant S I — G — 1 2 l - r e i who n i This usa’ e h a s i - - -n US gr aup u ar i a - - -- --- - - ann e ssa j - -- r ’ v cc- it e Jrsr ssod n c — r n t l ’ ’ c r e s e r c h have been uth cr Lsca to o - c AkiUO i fr tnc U S rur inc 1 9 7 1 omer x C c e l i s t t h e papers I t oroscnt r i c a s m a d e - t a o am an In 3 wmo Irratitution a tn a t I t t 5 n t i l d r a iumid a t m ’a -t UrICIarnneL C L ir na i e ccn t ccprese s n a i a m on ttl ’ C Cce aetnir ‘ i r r if l cj on tue you - CarrU -LE t ut ‘i Broe frc r- esc r c m cr l - an Chapter o N jo th Kirs n i cont inu e - C its a-i repr e sen t ative from E T E L et i - a r C - nhtc- r r -oi d irectl y tee policy - - mci one ron c t h e d m r e c r r e r e s e n t a t i n ’ s of ’rk Plannina ’ P’i c-n t r Gcc er ng Corror i ttee n s -ado it possr b le to discuss more ti-c ly e t h e prnc ’s ai mac proolems raise-i Pc’ t a project t ars ct is possible to rcno r ’e lc ’ v which uses of A R P AN l ’T f r cn ’s he hc ’c r c m US Are pec Ls b -a lj i l l c at ia -ms w e r e refused wh cr r n c ’ r l c ’o C use of Akl ’-k m pnn 1 0 t i nt resi I a t in Ln e CS or i n wh ich the r i c ip a i ii i g u LI I - - a i -i for a r i e c s i q c i traff i c It has also been uo ssiLle durin g 1977 o a i s c us s -rr c c-n t s i c w i - i c - h w ill be adopt c- i r o t a e r ces” eccial service is expected to become ce r-it io r ol be ’ en t b • cr w no ea - i v - -- j - p ermitted to ca nt nue u ii U -- m n k l US -and tia c’ ti- it C US As cm m a t t e r of poi r cy an y caL a c os mcrcim icC of m e 1 sec-v i c e w Lii not be ARPAhE link - 2ER 2 S 5N ’v a -” e t 1 i f l tL a J i i C i i i 1 The UCL h iS always cevc-Ic ’ tnt ether AR 1’ A h h h ost s it e s i-i Ut ’ m c e c av e i J fl i f cant lv hid t o’ ’or ae eci s cm l mc r e - a m l r q - U C lA t ’ r er e te s y sa c s - i-a 1u f the n -c - r of 1 2 we s h ow be Iii’YA •n O 1 i 5 am a’ t i c - t a m r e n t vca c - r oy a l S ic j -n c i i at -ri tao ta rget s n i t - tar a nerf- rm anc n and t P D P ca - f isr u - mm ’ - ‘ - -i i om onto •s v s t c m ir e t ame - s- U or — Frrr- L n i v e r t ’z of ‘ u l a m a m ICL System 4 7 2 - in Sent ren a 2 4 amount of gL A t rot-i Gc rdnr move required t flo — 2 7 c ncernr-C s N- -s AtE r e c t i i r e d to im p r o v e its AO’A 1 his an Ang e-s t t- trio a p e r t i n c a c t i v t” is d i n - c r s c e 1 im ra t -ra c ry e is a nd n m r m e d e r -t o rn C A i nt er r -rjscusseo a s t sf i n t e r fa c e sy s te m Se c ti o n 2 3 One t- -oceciurc ’ o natne traffic and all an ose- un v e r r c m 1 ones l e d Sw I T C H i s d i s ss -n t i r m q in tue -“s- cc Wh i iO mr - i the C h ap t e r m ost software to run under this ci r - -ter t r on Durin t 1 76 th ere li e bee- n considerable d ---vrl op ent nunf m q u r a t ion 4o8 ’ - -c ’ r - c m t t - hcc a t ar -u - i h i s-a c front—e ra rr ’ c n ’ we h a ve m a - r e C O f l S i s c-nt m easure— - n i d t - 1e n d - - t cy St O S ‘n e U C L A m f r r 2 2 d EC r mt e t y - p- i 3 the or o qr e s ma re t h e SWITCH a ‘ f u r t a r o r s ’s ’s em s a c t iv i t ‘ i- ch i s t r e at e d it a T h i s mi c cn o i et t r ed e si of t h e f r o n t — e n d n of s ys er-is r r ot a c o l s a e v e l ep a c e n t t -’ - so U- i a - ’ - I ’ ’ - c - a tricre aco s I - - er •‘se 19 7 T has ’ s ‘‘‘ In lect ccc i a ‘ E ’ ’ c t od moo s o f t w a r e i n t h e -’- ’ A h’i’ pr ‘c o l s t o ones u t - s i r e d b y t h e card we nave e r v e r y i tt -tnt m ’ r r T t r - ct s-coa x i c-rn A h f m or e - A-i a mca -as t a ner t a of ‘ cr c e n t c h e ‘n-oran i ci ’ s1g’si ic’ 1rco t - UCL PDP - s it • r -- y e a r rt’ c r f o m n ab c E u t - a a r m a r ’ t r t e n e n e s ts t s de s c r h c a cast I’- ‘he 5 1 at ’ - n i t r a ‘f o e t he 1 -C t e tas ed-a St im t ri -are h o s ts r i - - ttacned 1- -ao’r r at 1 cndon Cc r’sai ter Ce n t r e CDC O 0 73OO t a The a t t a c - s m i un t -af w t h t 1 1 - - develo ceLi • r - s’ ’st e i ’c h a s rec _ tc- riosas are am ‘U-is r -l c tori s o f t --a -i nc - fm n r r a i n r f i c a n t h o s t •ce rip’ at -r ‘o - er r - -a t — e a ’r remote ou r s e l v e s Pot h v ’ c -r vcmI IS e t h e a - - w r c a nests a n i S a g a t a t ii f t c r e n t rr t crt ant cr ’ caq c irac r e in the UCL -ma the move of our i n to t h e new Cower S t r e e t p r e m i s e s This ‘ instruction of a dup l icate post office facility and the move of cli the grouD ’ s - -orn uters l’r’ris move was achieved —4 — 6— ARPANET SATNET iT N _ RSRE POP 11 _ London 3 95 PDP 9 PDP 9 H _ _ _ c — r iq 2 1 6OOG7OOO A RPA _ Stutt gart Blacknest Locat term i na ls _ _ _ _ _ _ _ ULCC Cuiham ICL Dote 600 I line Datel 200 4 t i nes U CL Cea - f i u r at i o n 1 1 72 _ _ _ _ _ _ _ at t h e end of U- 76 J - --- ‘ -- - 48 a -- eo u st 19 7 t ‘la r e ‘lIP was - r o e t iL nc ’e ar r d n c o f f t h e air for a’- out o at i v c m e p c r i a ’a s i v oved w itam flu1 fl IOLi Ju b o r i ng The ncv erscnt of t he CCL PD s was a l s o f a i r l y a t f o r w rr r w i t h an u n a v a il a b a 1 i t ’ of orm ly about 1 w e e K st r a i j m The h u t on e of second I- CL P aP 9 hecam e sn r -r e wh at damaged in t r a n s i t and gave c o n s i d e r a b l e tr o i le ove r a f i v e week p e r i o d o c r e i l n a c - our a e s e a r ci ’ md s e r v i c e commit” only oat’ Piui 9 fur c- e us to r e - d c e abe u v a l l a b a h a y The d i f f i cu cv ments w i t h O’ i of p e r -ecu ot t h e Rb and R SPU’ m a c hi n e s d u r i n g t a i l s a su b s t a n ti a l r c o u c L in of u s a g e of t h e s e two i ac-ai m es r e s u l t f t ’ ora a rr r o p o r ti o na l sh a r p e r th c a i-i - a i d T r i s r e s u l t e d in — far r e d uc t i o n with hours of p r im e t i m e •iv i f i b i I n t l T h i s urounit a rc’ to cu again the need to p r o v n d e u s c i 2 4 — a m o u r a v i i l a b i l i ty ‘ui poss t U b f o r any h o s t s o f f er i n g A t ’ i rd s er ” n c c r s g a t e w a y b tv -e e n t h e p -c- e a s c c e ll was i n s t a ll e d a l s o in e an i l Du r i n g t h e c on s t t u et son of c h an - wore a n umber cU - c-s’-r p -ut€ -r te n etsa rK came p s e v io u s y o u r a r e t h a t so t a ct t’ e C— t r i p - I c at i o n f a c i l i t y made ia cam e CCL c or r su n i c at i o n e qu i p m e n t a t - a i r c h a n g e s over SATNET a n d ARPANET t h e C o o l ic -ate cant-c-e r-I The con f l e u r a t i o n f u r t h e r hn - t to a c t as a 5r d was t r o v e a w i t h l i t t l e d i f f i c u l ty July t t h e er b of PDP1 1 wa s in s t a l l - a - c a P1 P 1 1 1976 is s h o w n it Fig 2 1 The t h e SATIC-If g a t e w a y t h e T I P h a s now no room f o r cmv sr-er A i d c - c i 1-es igrc C e n t r e A t l a s c o n n e c t i o n was r e m v u aiaifl a ’ - c c - m i s c t h e At as r t s e l f - ‘cc caoi ng sat o f s e rv i c e Tne RL anna RSei c omp u t e r - we re a t t a c h e d to Pi P 9 A and w e r e in a l m o s t cor t a n n u l s e r vi c e t m r e q l c - i t the y o u r N r e ULCC Cu b rcmm - in C EPSS r e m a n m t C eT h er r men t’-m i a ct i v i t i es f o r e a t t a c t m - s m to PDP9 iA throu gaaOA t n ot - onerous 2 1 we also exe erciecI ou r d e t ai n g i v e n i n F i q e q u i p m e n t so t h a t PDP9A an d PDP9 E f r o m the corrnr -a m c c ti c n c v e e s s L - r r t i a i i y i d e n t i c a l f a c i l i tt e s The nu cl ar of is Al thoucj h i t t h e y e a r ar-el w e r e there- on t h e n u o l i c sw t c he d tele mhone network was reduces are of our ma jor user projects an e x p e r im e n t of MEDLIN J -l mon et-i s under B r i t h i i i ar a r y a u s p i c e s was t ’r m in a t e d d u r in c th e y e a r or n e t - w ” r O r oar - c r c m e nt s i n di c at eu t h a t t h e r e d u c t i o n of ports on t h i - PS 1 N w o u ld r o u u c e costs with a n e gl i ’-j i ble i m p a c t t cr m n a i s on port avai iabil it to s t -m r t the i n c f l c j to users a’ cn t ac t - ca yc a t i v o r _ _ _ F i nally we w ished in t o a p o s it i o n u n d e r w h i c h t e r a -a i n a l be s u p p o r t e d n ’ e ir l y t h r o u g h -U ’SS _ the usage of ARPANET a r s i de r a b iv _ _ _ Ifo r — ‘ -- tb r ou q t RL i- m a s t s access w o u l d t h e s e r e a s o n s we reduced - -- -— -— —8— a r c - ’ l f i c a n tb v sw i t c hed the number of t e rm i r n u p o r t s o f f e r e d or t am e p u r in c ’ telephone network O p e r a t i n g S y st e m M e a s u r e m e n t and I m p r o v e m e n t 2 3 To support ou r communications software and its development we During the year we nave our own o p e rat i n g s y s t e m f o r th e PDP9 have devoted some effort to a cons i deration of tire efficiency of t h e r e s i d e n t s e c t i o n s of t h i s o p e r a t i n g s y s t e m and its core alloca” t i o n segment l o a d i n g segment deletion and t rsk scheduling func— tnons There has been little change to the input output functions dr iv e r s hav e been add ed f o r n ew dev ices as nece ss ar y and extra queueing was prov ide d t o perm it synch ro nous runn in g of a synchronous f o r m a t character devices at medium speeds A study by an MSc student in 1975 of the ARPANET-R L system running at that time suggested that marked degradation would occur i f core s p ace were redu ce d As a resr mlt of this work and some theoretical notes Hinchley 976A and Higginson 1976F on segment loading overheads and tneir effects an MSc project was commissioned in 1976 to investigate the current segment ‘oadinq nd core allocation ar deI tion algorithms and to develop new algorithms as appropr iate The report Johnson 1976 describes the extensive measurements of various types of user system s including the ARPANET -Rb real time system in norma l situations and with varying degrees of core restriction The object of running core—restricted systems is to simulate the effectsof a larger system ie the mul ti—machine system in the full core size that i The results showed The segment deletion algorithm could be improved ii Desiqru of user system structures could be improved to lessen opera ting sys tem o v e r h e ads iii The segment loading time s measured were a third higher than iv predicted in the theoretical notes Two s p e c i f i c f e a t u r e s of i n t e r n a l o p e r a t i o n of the op e r a t i n g system were adding largely unneces sary overh eads _ _ _ _ _ --- - - -—— - - -- — - - - - —9— Co r r e c t ion of isolated tm last of these is trivial naw they r a’-e As part of improving segment deletion b eer which is the way the operating system regains core space for current activities the core alloca tion routines were rewritten This was also necessary because of known flaws i n the curre nt alloca ti on routines A system incorporating the new routines will be tested during the first quarter of 1977 be studied further in 1977 Results ii and iii will It also became apparent during 1976 that the operational task-scheduling routine was not as efficient as it might be and a long term solution to this problem is being studied Howeve r two ‘ stop gap ’ measures have been introduced to increase the efficiency of the current methods Most tasks in the system loop on the l OOms c ioc wh ile waiting for work to do and this meant that an average SOms delay was encountered when work became available The cloc k interval was re d ef ine d tc be the m i n i mum of 0 aims real time and 1 n- s idle time thereby reduc ing the normal average delay to 0 5 ii cu t not overloading the system with if it had other work to do unproductive looping Another long term problem is the non- reentrancy of PD P9 code - in p a r t ic ul a r th is has limited access to global var iables and irter segment calls in such a way that the normal method of bu ilding a system is to define a ‘main task’ wh i ch does most of the work and only use independent tasks in a special situations within totally resident code ‘ Seriall y re—useable ’ global access and intersegment call code has now been developed and is to be added to the multi-machine system We d o not regard further sys tems develo pmen t o f the PDP 9 as a research activity and will devote as little further effort in this direction as possible Howeve r it is clear that the multi-machine configuration of this chapter will make new demands on the capab ilities of the PDP9s Our efforts will be limited to those items which are necessary to the system succeeding in its new role 2 4 The Attachment of the Rutherford Laboratory BM 360 J95 During 1976 there has been almost no change in the operational software for the connection of the Ru herford Laboratory computer to ARPANET The Help facilities and PDP9 response messages to users - - - - - - _ --— - -- _ _ - ---- I- _ _ -— - --- or ‘h PAh f have i m p r o v e d u n n a bo t h the f a ct l i tr e s f o r a n d e x te nt of sa- e of she FL system have in c r e a s ee ‘la n e -Is su e c n f i c p r o b l en s in t h e fn so t r a n s f e r area w e r e O l s o r e m e d n e d F o r ex a m n b e t h e p r i o r i t y in t h e r e s p o n s e t i m e ‘was ecurovod amy u s i n q 360 an d or e m a t u r e t l r i c o ut of sy s t em level l o n g c en n e c ti o n periods was erad icate-cr trai s rnd arisen daring t n e - t r a n s f e r of large files ama - -‘ utput ‘erv f r-er- su r s tt t ed j o b s A r qor effo aras gone e a t O dev l op i crc- a n e w v r s i on of t h e A Rl- -A N E T 3 6 0 s o f t w a re to r u n un -h er S W I T C H Sect on 3 T h i s has tnr e e components A It pAhi-f F Networ k Co n t r I Fi r st i t h ’ s be en n e c e s s ar y to sr - s l i t t h e P r o qr a m N C r ‘360 c a c k n a c i n s u c h a -way hu -i there are very distinct com n e nt s i n te r f a c in q between ARPANET and SWITCH Second a 360 h A SP c o m pon e n t h r i- au t o be intert a ut faced between SWIl O li and t h e HASP i nt e r a c t i v e t e r m i n a l d r i v e r s T h i r d i t h a s ameen n e c e s s a r y t o b r e a k Ut the transfer protocol f il e One c o n n e c t -on is b e t w e en t h e S W I T C H module into similar p-arts ’ h ’ -Ih Ahi- C NCP The second trees b e t w e e n FTP package and t h e SW ITCH and FTP and a HASP R3E p - a c k e5e have been required f o r CCL al l PDP 9 b l o c k d i ’r ’ ranrs of t h e o t ’r e r S i na ’c ve ry similar exercis e-s cornput er a t t a c h e d t o t h e t h e c hu ar g e s -wmi c r h a v e be en r e q u i r e d i r the A R P A N E T 3 6 0 s o f t - w a r e are -c iven in Section 3 2 ALl this software t i e exceact ion of that r e c u r r e d with t r a n s f e r has been w r i t t e n in n l a rg e ly com i s si c - ne d for f i l e y o st of t a ut r e q u i r e d f o r file transfer h - m s al s o - r — een w r i t tt-- am ‘ c- have a r n e a ’s t I c o pp o r t ’u n st y to pr - v n u o r - 5Uc - i5 b e t t e r s t at u s i n f a - r m a t o n t o r em ot e users on toe file traamr -i r - r -atse rat son s is n o t a A i t h o u c a m t h i s s o f t -w a r e - a un t s -we a n t i c i p a t e it system by the cr-r d of 2 The Att o -v a - a--n t necteir tI t n c ur l y d i S O no of m n -n c t i ve t e r r l n i l i a s t al l o -a on t h e o p e r a t i o n a l t a m e f i r s t q au rte r cl-h F - 108C ‘ n c n- - ‘4th i t i v a nc e as t h e ‘- f 1977 t i r e i ovsr L Signals arC Radar Estai lishment in C R Iot l-lbt a t s t - it R s i - b i b ‘w is co n— 1976 to ARPANET t o r o u c - a i t u O t h e c on n e c t ion 1s cluster ot several interactive terminals and a ve ry A co u r s e w a r r u n s’i m ” c e s s f u l l y by the S l r - p iC p r o t o c o l is u s - r sim u l a t e s SR C s t a f f a in W a s h i n g t o n w i t h US D e p ar t m e n t o f Oufl S w i t h i n am w o k --- ‘ if - - the - - - - - D e f e n se D 0D I partici— line between RS RE and UCL coming -- - -- -- - - - - - - ---- -- - - — - ---—- - - - ‘ —11— r it e t i m e t h e ecs- a i c e t ion ‘ us ic- t h r n u am f o r s i r r u l t u n e c - u s usa e ‘- - - - r e a v u r l ab i e At t h a t 5 cOe- PDP — 9B a n n n i n e rn r t s Aft er t h e c o u r s e w a s f i n i s h e d t h e s o f t war e and t o e s’s r n e t i c s r ‘w te- l at e r g r a t e d i n t o t h e s y s t e m r u n n i n g cr-i PDP OA so t h c t of p o r t s was a l t - sr- ac-n r- re n ur ’ r er reduced to five the availab ilit y went u p Since t h a t time t h e system has been a v a i l a b l e f o r m e r e t o mn t w e n t y h o u r s a day to US partici pants in the US DoD High Leve l L a n g u a g e E v u l o n t i o n Project 11cr M aj e s t y on March 2b 19 76 t h e Queer o f f i c i al l y opened the L i n k b y sending a welcoming message t o so m e of toe US participants durinq a visit to RSRE The operation of the RSRE link during 1070 has been as an addition to the standard RL ARPANET system however a further version of our software to run RSRE as a connection under SWITCH has been commissioned We f o r se e no p r o b l e m -in continuinq to h a v e the RSRE computer availac-le when we go o er to tire multi-machine op erat ica under SWITCH at the end of the first quarter of 1 J77 2 6 The Attachment of the ULCC CDC 6000 7000 Complex The connection fri -n ’ UCL to ta re University of London Cc-ra -ater C n ’ re ULCC is a 2 4Kb ps line terminating on their CDC 6400 which can access the CDC 6600 and 7600 CDC 200 User Terminals Currently the PDP 9 s i m u l a t e s feur a 200UT is the CUC ha rdw ired R JE terminal and n o r m a l l y consists of a card reader line o r i n t e r and disp lay unit Howeve r the PDP9 software has the card reade r and the line printer permanently disabled on three of the four simulated 200UTs this still leaves the full interac tivr’ -acilities of I N T E R C Oy CDCA available to users INTERCO” is the portion of SCOPE the standard CDC operating s y s t em which deals with interactive term inals The fourth simu l ate -a 200UT is reservedfor file transfer t h i s utili zes a simul rt oc ’I card reader and line printer The PDP9 software is part of the multi-n etwork SWITCH system see Chapter 3 The protoco 1 used by the CDC is a simp le polling one The CDC j 40 act as a master and cyclicall y polls the four 200UT slaves The simp le nature of the protocol is due to the fact that CDC 200UTs are hardwired b e used which Ic-vices and hence no more intelligent protocol could The CDC 6400 sends an addressed message down the line is decoded by the software is discarded If not recognised the message Otherw jse the softw are responds ap p r op r i a t e l y ___ _ _ ____ __ _ _ _ _ _ _ —- ‘ - - - - CUe ros rc i CCI of Cn e ss d r ” u c o Lao go - o 1aL or tr an j a s ion f ruilu ro Camir fO a t e ’ s r Js eu o c O i -a L t tame error recover’ ’ 15 00 - 1 i L c t u c a - r e u e e r a n L s r or e r ece p t c-f CCC u ’ uu A sit L Ir u r i se ii’s 0 1 0 d a t a ‘sr t C i ’O O L5- 5 cue - -t i--i i ’ ’ -t W - th the cDT - Cm ti 200U io n o rail -‘ 00 cars t e l l oe t sr e r - ’ - amseous ‘ra c p c n n n in L ame u u L e x s i s I r e of s n i e - i n z i on - - i - i n pc i LOO aul te S requeot 1 carce a’sfar It LOt t ’ s s ire uS c tr c 0403 rae i’ L S p e w n r- p o i h i c r o rs’ I I n i p t s k c- r r 1 r u o e tic am ’ t - P O O L T L ’S OIE ’1 c ol T he 5C tL srsr nal rr sa sent p i -n c soon ram s- rerm iraci cor n furth cr l in a message that a ‘Li ’iore in In r e - ei L nc r- en t h e P lf- I’9 s y s t em one - L e - a s t f o ne re ‘r—s bee n 5 000 cr r i i n e -ca c t W l l l b e a so p s frec rr - -c of w r i c h is t h e I Tare Cr L 6 4 0 0 oat ti es wo e t o t - r n a i s t ü o iL f s u r o ± i r t e r r a l t e Tow-sC ya’- fr-ca -w h a t a re ci a in h- t the errcy ef p o l l 4 w 11 sep er u cr-s o c o Sa - a - c cc- - t- nt d amy e f L -ai or st-soc ret000l i s s n r h L - a o tra - - 6 4 0 0 w i I ccnt rsua f I v r e — t s - a u s i o A O s t i r e -so L o i c r e a s n o t l r a c ou t an-s flL s— u ’ r o e j U l u nc ITIlT la a further state whlca case t a r e t h i s s i tu a t i o n a s st a r t e d u p lt can 000r’°ra i so r ’r t T i t i 00 h o ’i f — d s o i t - x mon tam u s c icr rra arily for toe C — i 11 r to C hI C ‘ a i n f r - a n e - coca t h e r e c e i p t o f -d- t 0t ’ i s i-sn of c r c print e r o u t p u t - ss qe a r m up to 13 0 0 e s s ir a c t e r s I n ot or of ocl r0 r-ssc rn n e cc si z y ‘ s oce and z e r o t h e re a t a ’ c nc t r o d we h a v e a d o p te d o u t OOLJI for si-c ot y e f fi c i et c J t I s r t cr c c t i v e m e s s a g e s -a s c-n e n a - r e - oamr s pr o b - -ts tar s a i r cr c - a s ur c a o c n t s between - - - r S e ct era 8 5 h i- ’or sn o r t e CCC 0lI l a a ou I irrc - - LOS - S h0 ic ciorrection i t was p-oss b le messu ces on s- no on ’y ic r a i nar- pri r- - -r 001 of trio n e o -r L c PDii - ur -r o 1i h-p i n f o r ir ot r o n c o tr s m j 13 SrI -n e s a r - q -a 3— CLII u thou it- n I c f ’ irJ b cur s - 1 h - Of amamo at 2 4 G s os -or — g i a no am s e - -s n s-Icc- wer e t r’cIuc-heut s c st n of 9 8 nc-urn Tare -c c low fIos r -f are caused 7 0 ms- -P- - r em tc rr n ”und tjrrc bc-cause ovi - r Lr- - 1i is o h s CiSC p r o t oco ‘or- s h C t f — -_ a L a i ex f i r -c c u t of ev er y four asiossages are o oat of 37 S - s sa ouA am mr ocr -s r e - a s i c on t co-rtro r- a a- u - mua t am — e t r a i s s ru s s i or s a ’ad t h a t t I r e 6 4 0 0 r ’ ut c ar o c e t ro n g i v i - a h ey re so a l itO is _ w n y s r ca - tsr t i c pr e ss-co 1 v ca a-r oca ly des ir -aned 0 1 0 r- - - - ‘ ti -f at- u t — - L ’ C K ’ tl’s ’m t - m maximua - t inte ractive acm ’ as - t ame - of Cu L a woLchr 73amj s urL r e - - a son -aU 1”- C— — car am - 1s for us - with a contrc i1e -t operate at onl y a bou t f400bp s so f o r rn essaqe a of t h i- s l e n q t h is not - -‘ - - - --- -r--- --’---- ---- — By th e c-rid of 1 3— t he r e p o r t i n g per i ca th e CCC r- u c h i n e h a d ace- n i n t e r f a c e d to SWITCH and r u n e x p e rs o - e c t a l l y It over recentl y to La me b y t i n e m before a ou b l i c s e r vi c e CDC SWITC h A full staff at ULCO f o r t e s t tn q bc-en p a s s ed SEa S the The f i l e t r n n s f c - r s e c t i o n of is a t t e m p t e d and A LIP A N ET p r o g r a m s a r e coded b u t not y e t i n t o qr a c c - c l s e r v i c e i n c l u d i n g f i l e t r a n s f e r is sc h e d u l e d f o r t h e e-r d of the first quarter of 1977 St s h o u l d be ‘s ted t h a t a ll to r u n u n d e r P-WI T Ch se r v ic e w i l l RSRE RL a nd t h e CDC s o f t wa r e h a s been d e s s e r r e d in t h e mu i t t — m a c h i n e e n v i r o nm e n t sat b c T he coiam crnce o n l y wa ’r en t h e c omp iet o r - ui t i -ra’iocosi ne r -ricdin c LL C C are available for very subc ant ioi periods an t h e same P D P 9 7 The A t t a c h m e n t of Sof tware f o r t e CI LhI A C t i c- c o n n e c t s - s n of ICL S y s tem 4 72 t h e Cu I— ’b Oy s t er s hOt - ois 4 7 At t h a t t r’ e i t 3 0- of c i I i r st e i n 1 7 - A was stated that for terminal access u s l i r t r a c - d i f i cat o rs w a s Durincj toe oust year the s y at o - h a s b e e r r u n a ir - r e s t implemented des c r i b e d in S ect i o n w i t h o u t a c ha r ’aq- i n sa st - m d— - a h o n e vers t ora m u l t i — m a c h i n e s v s t e r- - is s t r a ig h t — f c r - ’ r r u r- - rogression to the The l ow e r level l i n e and b l o c k p r o g r a m s were converted in t o t h e monTe r e Cu ir t’sd Cv SWITCH w he n t h e y w e re for t e s t i n g incorporated into the curi-en t systor-’ u sed ‘T’hc interface level progr arrrr - Isesoc - used are st il the R L A N P N f - I T ones and we a r e w a i t i n c - f a r t h e RL 360 f a c e to necome o p e r a t io n a l b e 0 o r e ‘-ic g e n e ra t e a v e r s i o n of p r o g r a m s f o r u s e by t h e C u l h a m s y s t e m is a m o d i f i c a t i o n of SW I T C h inter- ohso so S i n ce t h e l a t t e r s y s t em the RU system we a n t i c i p a t e ve ry l i t t l e d e l a y b e f o r e C u i ham is a l s o a v a i l a b l e on t i e m u l t i — m a c h i n e sv s ter c - r a e t h i s is i n u s e of In i e Kirstein a t t h e end of t h e f i r s t c o ur t e r 1977 1 976A we mentioned that t h e file transfer would probably be im p lemented using modification of the 1II’SS b a s i c f I l e transfer protocol However C u r i n g 1976 t h e s ta s r l cr d i s at i o n activ it ies by t h e PTTs h a v e r e a c h e d su c h a p o i n t t h a t EP OS is c l ea r l y only qein c- to no - t s r a n s i t i o s r r ’ standard t uile we ourselves are continuing - i -ti vc - lv with EPSO particip ation i t is i m p r o b a bl e th a t C u lh am w i l l see C h i -p t er C ’ al s o Cc-come EPSS p ar t i c i p a n t s I _ - For tam s reason we have- r -c 3re -d to ‘r - cu i es cnc f r e t ’ r n n s fe r C- s Sr - dSS uy g i v i n g t h e m ac c e ss to - ‘ ‘i - ra — IC coma as- s t i s rU t t i r j t a m e u v p r ’- o r i at e camc ’ nnds - e done by CISC recoros Higginson 1976 The des sjn f o r SWITCal FTP was C sor- w - tr -e -r r I made sufficiently flex bl that the commands t- ’ - root p r- - r o s c - a s i came f r o m a n o t a s er sruc i in e P a r t l y b e c a u s e of Cc I Om s’ run ass ansI Oausti y b e c a u s e of our n-rn s y s t en - cas - v el o p m en tsm I liI c cor- y taut tan c ’clrraa s u s e end of t ne secor’ru qu a r t e r of s c n e du l e it t r a n s f e r - si l l be op e r at i ’ ool as si ’ — t e f o ro o h - c 1-i77 — --——-— - --- --— ‘ --- - — -- - -- --- 3 CI LAP’i’io- 3 1 S ’CITCH— - S A NSIWO s I C- tF c A i d Introduction s ort ie s bo Lts tsr lPL v r-tual csI is b e t w e e n n e t w o r k s and - as am n e t w o r k ac ce s S i n t e r f a c e tC w n i c h h o s t s Th e UCL OWIT C aI i m p l e m e n tat i o n to n e t w o r k s v s ma or int - sm r -e’i a aa -y ‘cori plater The a p p l i c a t i on to t h e i n t er c o n ne - s t o’r is of nem - o - r k s is d i s c c s — ea i in d e ta i l c - sirI be - ‘at t a c h ed in Chapter 4 I n t h i s s e c t i o n we d i s c u s s c e r s a n c a t i o n c o n s i d e r i ng b e t a s i n t e r a c t i ve t e r r s ’ i n o i data transfer by u s e of l ev e l of ac c e s s f i l e s e r - m a c f a r protu’s c-s s a t t a c h m e n t may be by e s t a b l i a - s i s c ’- d - -oan s s t at d o a h - a n t s Un ansi b u l k Trae rsost o m o r o s o - i s as the 3 60 HASP s t a t i o n or by a s p e c i f i c a m l l y — c i e s i q s - ’d C a t s i cxchange mechanism Reasons for interposing cormnia s r an in tcrra c-ai arv may so - 3 r e at f o r c o m f o r t or alternatively the core occupied b y f o i l n et wc ’ r -s p r o t o c ol s vary The a d d i t i o n a l l o a d to t h e m a i n C I U osca y h e Ior e f s e sm — i t lv t h e pro d cc is t h e r e l u c t a n c e w i t h good r e a s o n of t i me sy s t e m s s t a f f t o p r o v i de and l a t e r m a in t a in tidy be too g r e -u t new complex s o f t- w a r e T h e r e is a p a r a d ox n o d u l e s b u r i ed here deep in t h e o p e r a t i n g s y s t em One r e a s o n f o r t h e s i z e and com u l e x it y of t h e h o s t p r o t o c o l is t h e i n t e n t io n of p r o v i d i n g a very f l e x i b l e network interface- -abbe to support a v a r i e t y of app l i c a t i o n s offerang secure message delivery w i t h so pn i s t i cat e d c o d — t o — e n d fl’ w control by ru ar ov rn g the protocol support into a front-end mac i r1 ot- i nv a r i a b l y scsaie of t h e f l e x ib i l i ty may t e l o s t u n l e s s the f r o n t end hos t pro toc ol is i t s e l f made of some interest c-qu al c o m p l e x i t y I t is a m a tt e r of then as to w h e t h e r t h i s t e c h n i qu e of h o s t a t t a c h m e n t should he c o n s i d e r e d t h e norr i s -or w h e t h e r it should only be used in specific cases mis r s r e r s t i s n o d above is a t r e n d to d i s t eir o r t s- In some ways of course there some p r o c e s s i n g c a p ab i l i ty i n t o m i c r op r o c e s s o r s attached to a m ai n processor ansI sharing store w ith i t Even i n t h i s case we are i r m t c - r - e s t e c l in t h e a m o u n t to w h i c h n e t w o r k i n g p r o t o c o l s car— be contained with kn such a microprocessor function At UCL we have e x u e r ie r i ca c- now I BM H A S P sf -a raunuf act o r e r ’s f u l l - s u pJ e x a manufacturer ’s half— duplex protocol ad hoc asynchronous protocols one to a mainframe 200 UT protocol and two GEC 4080 and tIme c-caner to a m icrooroce- ssor I OS bb 8ObC no c i s r a t o r of te -s ri i i u s e r s tiS Cr s ’t b c i cm s er v i c e to networs u s e rs c on n e c t i o n s is d c s cr seed C x a a - - e ce n c e s Ifl t h e I O ’ 3a 0 r i a ’h OC s The i nd a v r- al urn pr sav -s h- tails of -s-se in s i s ap t e a and m u c r i of ris e a t t a c h i n g t h e 1131-I 360 to niI -A flI- w er e Ce s c r i cC in th e p r e v i o u s r -in a ’aoa l n c -C ar t 3 On± v K i r st e i r s 19 7 6A i n t e r a c t i v e T - r m in a l Cs-r vlce A ll the connec iions completed a ’ sICh h a y - ’- i ncas le vel or above t t ame - t ea ’s anal B e c a u s e m u lt i p l e ser v e r h o st oc t s s c r t s - ” s s r e no m e sas - o as osi exist s f o r r c b e c t i v - s u p p o r t e d a t ta m e sam e t i r-r e a n d o n w a r d a dd r es s i n a it iS r i e c e s c a m s ’- ’ to’ ask a S e n o i f - n s f ‘r s u s e r s t o t o e a e st i n a t i o n t a m e y sierra r e ott s o - s c a r I n c t - n P D P r se ace not a b l e t o add r e s s tame require-a sorv oe m a c h i r e Cs r a e - c r y fr - a t h o r s o u r c e m a c h i a s The e n h y p o s s a b Sc d i r o c t motamens c-aauld h o tC m estqnate a parti cuIr -r h o st as a n ur t i c u l a r ‘ b c - a t ’ s e c - - t o e on e x p l i cit r e q u e s t f o r c on n e ct i o n a b l o - sc posswo ao ‘wa - ic ’ i s n e c e s s a r y in SOar- of ‘she c o n n e c ti on s -wh i c h 3 2 1 Tes snr ’cr A d o it i C-ontrrol -‘-an csm an r oct ic ’ n -we U -r t ‘r C h a ra c t e r Ter in a l S e r v i c e i n get-icr r i we a r e c at e r i n g f o r u s e t h r o u g o O C S C ‘ - re -r c eisr -mru ’ter t er m i n a l acces I s t h e n o r m and w h e r e t h e r e - is r ot e c a r b i ne of € t ch ch -’m s t u ct er f r om serv sa e s t h e h o s t to tOss t€-r r i a n a l i r e- o r i e n t a te d W t l e a -y Ten ex to t h i s ap - r o a ch t h e - L-S- 30 0 f - a c aa’erl be operate-s on a line—at— r - a— t irc c- basis For users of t h e 00 t h e P b C -’echoing so the t e rm i n a l user the si ad - c a ne he w r u l a t 1 ues e x p e c t a lt h ouc r’h o n l y COdj ete lines ore se -la -Sc len - - - ‘m r Ce - - hr tis 360 in the reverse direct r ann thinct s are a-sore corraoi ic-a L och a u s e r r - t t E i c h O ’ to the 360 won d I s ke control characters for’w - rLSco cc- ‘ hs- n e t W i t t ’ a h- t carriuqe returns t h e casId It -ir a of a ccmr r i a l c rem r n wroald be — or e are nl as a f c ’l low i ng c o nt r o l c h a r a co c i oee n ‘s u o t ’- d t a sa t a-a ‘ l i ne ’ f r o m t h e 3 6 C Lt s e- i f sa ’a - m i n t c o fac t a p arti l line of centr eS c h a a - - m c t c r - s ar i d r et w i t h o u t c ar r i a g e ros of solving _ _ _ _ toe prob i an _ _ - _ _ - r 1’ms Th m s mc Serire tore-- el so --_ T re a an of - ‘ - a space se n t o u t s m a t 15 the pr tr- - s a t c c w av a r e a u s a ’mr - l c i v rat w a y cOfl VC flt i -T 5 O h ans r -vi i inq - --- - --- to to co w- - am - in S u t L a m C ate - t oe -r ic a - ny e c h o i n g it t mac a ae U 5 e u l - abi it y r c p l ceni i -sn-a I t e a a rmit - ir - a access pOiflt Higginson 1976D This CAD assembler approach carries with re-a - sot - ec a so sang s Is ice 1’ con ‘ -c-n a- c a t or a c- n ’wor s th oicket a s s cmb l e r d -- it the i m pl i c it i o n s of a cage—mode terminal rather thafl a duplex ch a x a c t e r - by - -c h a r a c t e r t e r m i n a l access such as Tenex b y t h i s approach — U s e o f h o s t s s u ch as the 360 is made easier t h e b u r d e n of o t h e r t h a n 1 i a ’ a o - e t — a — t i m e n s a n d i i n g is too much f o r m o s t networks of hosts to bo- mr l ’s remains t o be Seen wh e t h e r t h e los - of ch a r a c t e r access w i l l r e p r e s e n t a l a r q a burden to bear In the current set-up 360 users into AR PAN Lk ’ -race is-star l o c a l and rem ote echoing It is r ther cifficu lt in f r et—en d S i t u S t i C n S to exactl y t is e r icj ht ba l an c e of f c i l at i e s Ch - r ’ s-t c r — s e t m a p p i s is a n o t h e r t h o r n y area -i The t JLCC 6 4 0 0 f o r c a-s a mp l e o n l y aro s a 64 “j et c iaracter set access tlrrouq b’a 200 UT p r o t o c o ls Genuine 53200 t e r m i n a l s ‘- -a 3 pOrt two local h a r e w i r ed c o n t r ol c h a ra ct er s f o r ae l e t i o n of characters or i l f l O S control charact e rs Use of t h e CDC 6 4 0 0 i tse lf does n o t r e q u i r e There is no wcy we can support tar e U200 control c h a r a c t e r s e a s i l y w i t h o u t chang Iraq T E L N E T c o n v e n t i o n s f o r e a c h A RPANET access d e p e nu s ng on s tat e d d o st i n a t s e a This Would mean casanging these conventions actually during tise excrram a qes with the user T h i s is casir p lex due to our m u l t i p lex e d ccr ’ n o c t i on s of computers but w o u l d n o t be too unreasonable in am n o rm a l s t u a n c o r s w h e r e on i y one host is amen rag front—ended at a t in e our d e f au l t TELNET and to a low i e have recently changed convention to support two or tnrec e d i t i ng r- haracters S m La er c h a r a c t e r s to be p ss ou a 0CC m a c h i n e of t i sr o u q i s - h er tar e c o u r s e t h i s j u s t means that control c h a r a c t e r s need supress on before reaching the ULCC 3 2 2 A Single Service Level The t e r m i n a l - l e v e l m a p p i n g is close to a t r a n sp a r e n t m a p p i n g - e s p e c ia l l y as i n our erase most control chit s-actors ore passed straiqht t h r o u gh and we are not support ing some of he more sophisticated options of t h e A S C A N E T TELNET protocol The difference is s u f f i c i e n t l y sma l l to enable us to implement f te transfer protocols described in 3 3 wr i c h use the interactive terminal m app inci - - ‘ acc ss Ire b i ar r y a-mis s ‘15 ° i h Si- c- S a - - o t i o s in o a r t f o r Li i rr - ’ 0-Tat a Ci - ih - a aiirS 5 a e arl aT - o aLenuea at ‘east al U 5 i i iaa this case nor amy a-a fc- r ca h a-icC f i 1 5 t r a - a n s f f - s ’ p r o r c ’ a see ’ C h a p t e r 10 It appears toot a c-araannun soa -v - ce ic’rae a t t a r e w o u l d be store approasri te to ami I ro me - ra t -TaOU ’L i c a- t i o n s some form of c o n v e nt i o n for t e r a - a s s a l r -r-stocol a 5s 0 ang ‘was -a -c c o i l — le vel with ap p r pr L at e t ne--en s in eat hihether opt aan sh i OCUInS se -n ccmr’msr d-s to the PDP Ch a -t a s t e r - r a c e w - r ’ i a n g e x e o r l u r r e e s - Ill a-costs - or a so t ao nr an oms on t t a r e - ma e- are -n t las Ire- 4 us ant isiSO or roe a n - as t am e n e s t rr cr ’j i ’n b - a c l U ’ e- - cal I — to te a-- e-- ‘ e ’ Control a-arpan on of an ad nec p r ot o c o l f o r c r a n s l a i s a a -as ar t - - ’ees tame h 12 an t ne A T S L U 3 0 f o r in s ta nc e his s forced taSO i SpoS t s e a of a -‘ I C r-ta coi f o r t h i s c o n n e c t i o n an ’- suer flaw control ia-a s e t h i s i congo -e o n l-’l a-w -as-r- a-ratly held a n the- PD ° ‘wo u l d be useful here t o t h e RL T - l-i 31 -3 maci in ana - f e r aapr oam ch fi cram s -- f e - i’ a r a ass el -ratio n s a c - I C - V- r ol -ean e a m in a p ur o p r L Ta t - 0 t t C 51t e -x c a -- h an 1 _ _ _ e- am 1k - - _ c - ‘ - “ C - - - t La a- a m Lee this shir v miii L hm e-cno r 1 in m o r e a’e a a rot iTacu - ‘ t-na- - ‘Iris CLaii ha -ac ‘ ‘ ° r-o 51Cr ch rot a- 5 - 0 3ote and tao tO - at t - t t i ’s it 50 f il e i t ra r st er a - tL a bc ar a i c- - C 5 - cO e ‘te am I fe- n t a r conocet son u r o t cO v i i l am - a - a r - s l ts g 5- c C L 3o 4 • 3 a hea St - - in c r m s e - -ua- O L ie - t o t S la-EGG 6- 00 u se Opt ‘ h-c s f - s- sea-s tr -t 50 ’ r ans t -r _ s-n an ta - e f ramao - -w-s r - - no a -e r e r e l f i l e -s on tne aS t e - r u u S t a r s - flCa - d r c ae ed i n i 1 9 76 we w i s s e d a - ed ias t aa t aae an s o t - L i e - u I r cotrn cs’ 5r a a-a tea ‘a ct v c’ I no’ “CSPANC’i’ ar During a’ cua—e Lu a Iso t i e a a t - -r f m mg a n e-’t a r e - a _ 1 t ‘- ‘el u s e no’ 3 the ro r c 110 not tr sa- e- am— - anne ha ’ SP - ’t c l Lu - 1 0 - sCar f i n - - - sto re ca - s 5 ’ OO C S iSO ac an d e r ’ r r s l e d Sn uie ’- ’ieus - ra s - a-ac s o a r s o fer s er v o ’ - - f i le c l s hoc e a g 0 r Sa -U l b ckn c o n tr o l a s a s e a n e am in t r an s i t is -s oc St- i ’ Ca n o -O c r o s s tc-Cti Ofla n c- a Sc -r v i cc s r a-p I e m c n ’ t s She ac m es - T ran s f er 6O ‘- Fire - i ‘ 3 3 f ar e a-se be i n s a s - d o of control exist ra c u r re n t l y tc- o r e v en t C’ - a n - dOn - C COG — ‘- ‘ - 1- rr 1 0 i a - L a I r santa-r n f i l e t i an s fe r o p t i o n Our model of use in the past has been to implement server file transfer only where the request for the transfer is initiated outs ide the PD P9 say in ARPANET The internal file transfer interface makes r e q u e s t s f o r a t r a n s f e r a c r o s s t h e i n t e r n a l i n t e r fa c e Thus requests can be made in t h e p r e v i o u s way by mapping the ARPANET file transfer rec ues 4- o n t o t h e - i n t e r n a l f o r m a t or by i n t r o d u c i n g a b c-a l u s e r i n t e ra c t i v e i e --ce of sot Lw ar e j b i c Lu f o r m u l a t e f i l e transfer requests The ARPAN1Cs FTP defi ant ’s a ama --er interface to its p r o t o c o l i n t h e f i r s t i n s t a n c e it t u r n s o u t to be more c o n v e n i e n t and useful to implement t h i s p r o t o c o l and map i t on to the in t e r n a l format The f i l e - t r a n s f e r se ice to ’ t h e 360 h a s f o r -a long t ime been one of t h e l e a s t r e l i a bl e s oft w ar e s e r v i c e s r- ’e h a v e p r o v i d e d It is still not tot ill y clear -ah y we hav e b ee n so u nsucces sf u l in g i v i n g a rd j e b l e s e r v i c e t ari a mu t ea - ily t t s - f r e t t h a t t h e 360 although i ce- king like an ir tc-ractivt machine -at c h and in t e rf ac e ELECTRIC is fundamentall- the file transfer ama pi’ c a ch f o r AR PANET and the new PDP9 is f u n d a m e n t a l l y i n t e ra c t i v e l i e s at t h e root of the problem M-e r- -o v o r fi le transfer does rei’a u i re a level of service of higher r e l i a b i li t y tb-irs interactive use and in t h i s particular service we - ‘i re seeing the limits of the front—end approach or at least our ability to write code which is satisfactory for this situation It will be interesting to see i-f we fare better on the ULCC file transfer scheduled to be in use by the second quarter of 1977 this time we are d e v o t i ng emp loyed on achieving t as much softw a re e f f o r t reli ab le connection Certainly as can be usefully The CtJLHAM p lnn where the f il e t r a n s f e r i n s t r u c t i o n s are obeyed at their site will also be a useful p o i n t e r us to where the we i - nesses of the approach rrhe prospects of job transfer protocols for networks front— l ie en d ed in this way are not high unless we can achieve better results t h a n so f a r i n d i c a t e d — 3 3 1 The I-a is m aetur I t _ — tan t s r fa c-- in D eL a il 7 s m - aa t i one-d e r l se a t l a - c S W I i ’ C i i l i p U S C S t h e basic data transfer faci litiLa S r i r e - t i dy presen t n S kITCH swIrcH addre-ssing -san a ls o t i m e - r e f o r e - r t - used An initial r eg ’ a m - s t is made to the destination FTP service- raa - i nq am 0Ic 1TC1I CONECT ‘ ‘hr i s f a c il i t y was impleinent ed with a service fac d us in op t j c - ara l pa r asarete r a n d i t is t h i s f i el d which is used to th e d a t a j r d i n r a t e FTP c o n n e c ti - n ir rii-iicit ly service Uavirrt estabi ished contact up is r -c-ser ” - i f o r e l a t e r and a passing mec saa n lsm is now empbc-1-oai as a comm and oa t h - -r ay be tween s u i f er a-he two i’IP modules aase- an 8 b i t b y t e as Commands a-he b a s i c w or k i n g Commands a o vari d - le -- leraqt -option strings an d r e p l ies to a c auriaies d r e qu e st at-’ t rm im a-at d s a a ’ a -s-at ’ s-oa r -a d ‘l I re c o n t r o l quantity th us osci h do n o t a t e - s UL - twc ’e-ri across C C a-i1 -U t L t i in a-i i of t- a -h i’ p t n e w ce o m an d r e c ru e s t s h is protocol f u n t i o n s 1 iflhiC to Ct i hi an i t i sal i n k t h e - m a c o i l i s i o n d ea d l o c k s i t u - t i - cc of - lais of -a r ia f e - n ifter c o mp l et l en o p t s - en s ‘ h e t i- a n s fc r st srt — r ca - T a rn i f l - J rs I t can be s i g n a l l e d f rom may be rL- ui li i tire or ig inat r or s k- s t i n a t i o n of the file ABORT I - 1A fom t - -e a se s C a s e- l u r ’ t i ae r w i t h c -Ut END O F dta - a i a cl ude t h e 5 o l b o w i n g sf u n t o A T E A C C E P T aa a -t e s t the- be- i’ sol - e- The commands a v a i l a b le FILE in comm un i- - P l a m i s may o c c u r a n d a n us t I i T 1 FTP - h a v e - to be c o l i c it e c i -wi tr i ri t h e PDP9 operat tw - Ad D sTo Rh D ir e t he meSra-1lc S f rom the sender and receiver re spec tive ly after a successful transfer A subst r ile -t u r t - initial option s in - f comm in d s ilso exists m a i n l y to d e a l w L t h up the transfcr as - t t i ii - include whet rer pie-v a ous v r rs ions o f directory should be’ de leted if pp nded The p o s s i b le o p t i o n s this f i lenam e in the destination found whe th e r t h e f i l e s h o u l d be to an -xisti n g rile or whet ti et the file should be treated as a j o b to be sat h n i i tt r e d L a the h o s t s yst e m f ir be run prc i 1 c I i Ic attributes ne -d t — -- - - - - - am i s- at cn l r rran a i ca t e-d such as n a m e — a cc o unt p i a - r w record l ’ r a s p ma rd of d i r e c t o r y m x t m u m within tu e d e s tL n a t i on — to file Size the I t ie ref as - a O a-- - X i O’ The- f i e x i b i h i t - 1’ of opt a usa- s am usc- t h e t r a n s f e r if it is u n a b l e to S of the required feature The a b i l i ty to use t h i s st r e c t u r c - f o r A R P A h k T IBM 360 ULCC 6400 and EPSS will ba r- under investigation in earl y 1977 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ —_ _ _ _ --“-— --- - - _ r LIIAPIER 4 1 4 ‘ -- ‘ ‘ — - _ -a- Oi cO a- a S u r v e y and B a c kgr o u n d t h e p r e v i o u s annual report Rirsto-in 1 76A -we c-emrs -rast ed t h e a l t e r n a t i v e a p p r o a c h e s to c on n e c t i as n - t - ’ c t - k s ° ne a i r o s h is to have any i n t er c o n n e c t i n g “ g a l e -w cv ” co n n e c t ed t o e a c h In n e t w o r k so tire- 5 - a- n c o n t i n u e t a f rameworks an a l t e r na t i v e a p p r o a c h is to i n d i v i d ua l in t h e i r act gn c re- - i n d i v id u a l inc h e l p f u l t h e su b n c t c o n t r o l on an cnci- t ti -end I - s i s f ea t u r e s of - The first ai-ur oach requires complete protocol a appi n i at a iatcsuv m a r t l i t t l e e x t r a p r o t o c o l i n each h o s t to p e r t i t a n t e r n a r a w o r k i nrs — t- re second a pp ro a c h m a k e s l i t t l e d em a nd on ti re c at e w r but recuires a strong e n d - t o — e n d p r o t o c o l i n each h o s t Accord me L y we p u r s u e d reseam r-cla- 1 ject r o e s on b o t h f r o n t s On a r t i c 1 o a m t -J i n t h e j o i n t S t a n f o r d BL3 I CCL p r o cr a s- for the des gn de v e - sasun t t r iO srmeas mre-rr e - n t of an e n d— t a - —c- rd protocol TCP—descr i P t - - i n Ch m t r 5 - an a - m ems t a ’ a - - t h o r h a v e t h e one h a n d we constructed a - a t e w - ae raoapjcl ng T hi s m a p p i ng of v i r t me i c-- a u s b o t h host f r o n t — c - n d s t i - m areas arc d e s e -r i b s - d d e t a i ls of During L n t c i fac e networks a nd c o n t r i b u t ion It h a s a i d e d t h e i n t e r c o n n e c t i o n of the total solution In ‘a us e- i -ii s a s e- rt L at e r t e c hn i its t tare oI l a - - a -t the the access these ncractical k25 c l e f r r l c a l l n e t w o r k j ot - ma - ’ th e --rajor PTTs has made dar -v I up I a l e n t the eventua l solution n e t wor k s l e t in this cl a t c r ue ir a p r o v i d i n g st - n w o r e ‘1 i n Ch a ter 3 w a t h f CCL s yst - - ma - C h o ’ c-i fa r ’ d e f i n i t i o n CCI TT 1’ 7 C a s u b s t an t i a l i t - u rina l an d a - ie -- S L -r ancO aced ir a t e l y 5 t im e c ui -r e n f lam 7 r — a-d- l i C k if psc kot switched to t h e p r o b l e m of h a s not by i t s e l f p r o v i d e d 1 9 7 7 we e q-ect to m ak e use of the lessons le arned from bot 1 TCP an a l t h e m a p p i n g o f v i r L m a l call networks in it- a iC V j fl t - in r n t - -tne -L w o rk in g environment of w h i c h X25 is a funda— mas - an t a l a-inst t u e -n t I n v i e w of e n e -r q - r e - e of X2 5 ta t - is a standard for network a ceas w a - w i l l r e v i e w h a ’ re - fl y o u r -su rren t be f o r e d i s c u a s s in ’ o u r a d a i arrv c-nents i n 1 9 7 6 A n i d i s c u s s i o n of how to i n t e r c o n n e c t how tim e m div a dana I networks are la m m i t unanimously for a virtua l call solut i- n - ‘- - - a t t i t u d e s networks must start with fo r e t h e PTT s h a v e gone There will be further -- - -- - J — 23— j i e m t s a e- I I - ‘a aol ‘ v i e g h ju z i n 1 97 l i m p o r t an t ou t st r e s s Lu Access protocol ta i f by a - e r c-lit t ria L Is mao -we -v eil i isa ’ ‘s o y L -y c ir ce -ram c -aC e r a m - st a The sender is provi m e i — ‘itl is c o m p l e t e l y r c ’a b i e - rare ssag es c o u l d oc lo s t or du p l i c a m a 110W mae fu n c t i o n Sc cf r r o w e v e - r maca ‘ In i s m a y against tOe ccimaaparam t ively srna ’ p ie - wa r m cia s p e c i f i c a a i i a - a r e a c e-u i am - is n et a - h f f s c lt n e t ’ -- - - c -a- ira - L ad a ma- 55 q- a i d a - e- t - -- -cr-c 3 C a-s-a t o t mar i o Ta - t - a - -- o r aa - a Se- ’ue -ra --e e of i- P e- at SC- h e - t v i - s 3 ifl i e -i -a-a t i ’e o r m s i q i n c j P r o t o c o l P r o V e-c m sens e a - r I f a i lure -- car i s t i l l r - r a- n as- ‘a- trram t u s e r s W i i e n d — t o — e m ’ u pra t ot ‘ ‘a- it as control faais p r o v i d e -U the n e t w o rk ‘ It e s S ent a a f L3 is 0 eca- d el iv e r y c o n f i r m a t i o n e-nd m e s s a g e s a r e o c - l i v e r ec i to see a - a i e1 a I E g R -oL or 0j 1976 are a m u up t e d X2 5 is - a L SO Su t e f a r tue a- m a t e - r c a - s a m e-c ia-er of networks There - ir e s t i l l s e v e r a l o pt i e r a s w n a car r e -ca-i a gr e e - e t some u ct am i is sa-m cm ’r ca - S ca re - nm _ c nioc1 f i c a t a 0 a b -5 - t h e Lde- d at a terr ina I e u1 n ’m r a o n t of logIcal lin k r a 1 - t i a rt a i l o c c u r ira t h e ar -e s of e r ro r r e d c v c I C a nd s t a t u s r e ’er r i n q a - t a-ave s t i l l to see h ow w e l l t h e - s e s a - m j e e a - a are know of n t r e a t ee a a - -w s- r ra - - connec t e d -r e a-S L - c t Ia o l - e s - if t e n wi I a uvet c t r km aS ’ a -aoe t ra ae- a t ti me - cli m ’ a- ’ m t a i’ e - -i - w f’e rc- - r u i r e c i 1 1 n e ne p i - a 1er ra ca -f a Vi- ac ot t c e - a t e - e l ma - is ‘ arm ’- 1”I”Ps c ’ se t e rr a a t r ’ecp ar ‘ the a r ’ - c - ij lj ex saa - p l e Li at s e r V e -r a m l e a s t v e r y sub stara t o r a t -ri teti Fam e r e - s er v e - 5 e c r e c m l t tie -se w a r e - i t 01 - w a l e - I t - -- - _ rat t ’wc a tar erw ase s it of m r o t o c o l S to tO - r a d d r e ss in g a nt 1 gateways regions u m _— -—- -- - - - - — t a rt i h e - i r o p e r a t o r s shc’ an m appr o p r i a t - Ot - E -x aa apl ca-s ar e of -Ive ramcj e t r i f f a c in a- the to tnc PdiDN s n e t - w e e k s w i 11 not i s trana act a n t - i - a r c s s f i el d All is -pa r a - me - c m a-as c us ’ - ‘ - r s w i t h i -wet - p t - i c la - - i- s e - s fur irmee - a d e conmrec t i t at i n g -un raea it to t o re— -stea risk c a - J I a a a t i - n L u p t so n a n ca s e of f a e I i - cs no ro-cusaS ry t m set up y e t a n o t h er - k - a - a a p l e s a i 5 - f a c i l i t a - er a these- l i t -U tmae Ir a i’-’ i f i uc a l - w e - c m similar access to reserved fields a w ill SouR- net’ w r a-r S - - m i i - wa v a - o or a e I’Tk -Ns - altcrn at in ‘a t ’ re - n i r e t me e t w i r m a c u L maa - - O - i r is- sir a- it a st i v i t a - si n g l e mr - L ’ w o r s s a ma - a l rat ’ - f a r - t -i mb imp l e r a - en t at i o n a - “f t 5-o n g a L - -‘i - it S be -tV t e-fl m e - S w r r s la I act ic - - ea— t c - — -r rout in g c- eLW e -em of - ’ t la- i” e” the t o m ’ v i r t uam 1 c a l l t e c h r r rg’ cs trr ff c t - ‘rede -raa i i o a n t ar e I 515e rap i d f i u c t u a t r e r m s also where r c-re e-re- n et -wa - -r h - - - - - - -——_— - - — I S L a t — ta-i m a L P a- - S i- is - CeO -- ‘ ‘ - - -‘es - ama a tm- U’- ’ Is ca-t n L am ia taC t is ef I r e e carri e - ti sea- O ’ m o o ’ - t R a-m ’L i- am ici L t ar e i n c a i a - ar t na m lirea ‘a- sraa ’aa - as P mc Pta l t a m Le f ak- c-c iu a e a- f anuc am a -cc a as r - j u i r rt a- eaa- t S a t a- 1 c ’ f ca-a-ret-at at - a- 1 i ta’ vi th -ox a sting c yst s r s - it - Cl a r et c ’rr - ctaL t h a t a c a r a- -c-ty of tvpea o n e t - ’ ” rb - e - t a - l i s t t a r - a stsbstai ’m t l al pe -a’ n od ‘w i l l to a-ic ‘wa -tb a for s a s i r e -t e e -n t CC such ‘ a S s - h i s ‘snd s-s i t Ia- oaa h ot h e r ‘ t ’ n e t ’- ’- a ’ ka m- ’- ’ a - t a n j e to L a t e r c c f l r a e c t - p u n ’ s- a- m I t i a - I r m ’ n j - a - I - i t e- On t s-aer e w i l l Ira - ’ t a - ’ t - kis ti a V pi I ‘t e a t - -c ’s Ikamsall a i a t ’or a m i’ p r i s a ast a t ta-c’ Icve -’ L ra - asc ’ a ’tc ’ a c ’ i t u si a- ’ a l l at I h ‘ a-ar s i ma- rnma ny w a y s -a-- - h et sat a-’ re- a- a- - - - - a-k - i -v idA r’ a- -a 1 1 t n t cr f c ’ C - i r V e - r S O i t’ ‘- tlP ’ a - f i’i pr ’ v 1 iae - i - i a t a c a r a m s by a ii a-lam t a - a r i a - S t a c t t ce-ao l a - d a t e n a ml ci m t s — r a r r - ‘ocr ‘ a l I VLa -ta -La j s-it ch - ‘ ‘cs- e inc s- on -where n l a -r r i c d a 0010 s r i r ’ o m ’i l is i el it -t-t an t - ”o -s a- S i n a- - he - i iS I L s - nrrcsf’ - ” - r k art ect S ilt of As ‘“fi ’ n a ve we- ‘ a - r t - l e - a l l c e-r c a t m - n a t - d sra “ call I a - f - Th ma - - i r t mal fl en f o r each t s - c - e i s a - c-s n ta t he t ire- r e i n d i sc u s s i o n IRA I SPAC Is el I s arc t ’w mr et a - - i -r l -‘ ‘- a - - n e t am n I’ of a - em - a- a - d ma - i i n r e - r - r - s c- disr t - An sI - i ’c m s s -c ‘ ’S’ ’ e-a o ml t at i’- ‘s - c a s t 3Ck - Ool h -1 eves tIre - a- m t a m g r a m s subnets-a- - m a- Ia’ a- ’ i -aj a r - a tla -a - ’ ala - arO ’ I ta -a-a o u ch t tao - a-- m a- - l e - am a ’ s a r ama 1 C a 1 1 -‘ in a ‘a- t - a tee I ’ m - r s S ‘ o a k n a “- a I a he 5 a e - a - i I U - c a- - ‘ r t ’ - of a-i L a ’ ’ c 1 The t o r t - ‘Lia t a g r s-m a a a d ‘ a c r -l a -ac c t h e ‘ - a - c a - eras a the san la - iss a-a - ’ roal t e- a- s l en -a s t h a n a -n a - sm i me a mns ise-liovos a- ac l a s t ea - y da - - n n a -t ’ t i r e - a - 1 - a t I S a - e s o that - -ai r i ’h t a -h a -a - i n t c r c or r a— a r e to a k - rat - e li ton t a r ’ t - t a t ’ ua l c a l l 5-i t t -S- da - S C- 5 I r e - ad o d -- m cua st u aa rr s- r ra t i - rks Wt ta m wi l l - a r e o x a a - - i r I- ita grani rs r Sal t it a - mel - ’ n - ’ - a-- -’ ma - cat e n a t cd a a i ’ a - l i r a -- m a s s t h o s e t e r a f i n g on ‘a ’ s we w i l l a i s c U m a - s c on c a t e n a t e d n e - t v - - r h s ‘w h i r ’ i t is iro ppropr iat t ca a - mak e - ‘ amy a s su m o t i o n s a b o u t d e l r e ’e - r - or d e r r - 1 t i b i l i m ‘a- - ‘i du p i i - ’ i t ii cra - I n su c h n e t w o r k s -i y r r a a - i a a - — r c - i t a i m a e r a l t L i I C ’ l i t ew i ’S’ aS m o l t a ta m aths reconnec t i o n ets ar ‘Uta - L u 1 l y d i f i t - in ra e - I - — I m a tS S t n - ne - s c a r d — t o — e n d erie oc o i s en fbi - ‘a-- i’ a-and trace tori Liii ’ tIe Oer n - a - a - I - Cc ri -cl lie-re o u r ’ -n - t I n ’ f l c t ’r l t r -aa- a - i a - t ise -u i a - a - b a -t to tCid allSlfl r - IiN C ’a ’ sl I ’a - m isi - i t t - - r i i s en - It r -e l i t Liqt re-qu a iii- ‘i - - a ‘ -Ia m — t tI I ti in ’ mn i p i - —end 5 aleI e- ’ i - t o c - - tot- o ct - so f a a r t - -n ’e- rt ’ci t i n - ‘I a-h a-ave centred in ‘l a d t ’ t er o f Ci’ am a - t er 5 Z sa - ot -rm in l 75 It ’ at i-os i n i p a OI e ’-Ols to a - SIICC ’ j’’n 5 need a-a- ’ta - l 5a- ’e U ’ t o ve r - -- r c l e-t v e l y Ofl t o p of — --- - - - - —----- - S C ai — ‘ n e t w o rk e v r r t ‘a nal ca l l --- f aa-c- — --- -_ — - ‘- - LaC ’ a- S a 0 — - — -- -- - - 1i m5 ape -S - s3 nh — — - a- p r o c e d u r e l e au s to e x c e s s i v e- r o e -u r a m s a n c y ta - a c h c c i i na- and flew control both at the level of tar e-- s o f t-w a r e needing to res LdC inf u r ra mati -m n i n t -ae n tar e host sated n a - inc - a - m e - c r - a - ta- ca-t i ’ r m e - e r s ac in a u d i t i o n al ia-tea-Icr in s - r e c tr c e Cdi t e a - r a r e s -n a - n — t a - i l a - a - iCr C t w o r k a- nso ws a-a P o r e - - - um -t e- - - aml ’ i u a - r L a - C i t L a - -U a--c a- t c co n t r o l s can ue isa - a f t i c i e a a - t “ 0u t — o I ’ — o - r ’ a - - c - omai n a t i o n ’ v it ae t e e c- -w e r — l e v e l p rotocols t r e a t “ p r u a - e - c a- I ra - en r t c r e f lc c ’ r e su l t s s e e Chapter 5 ‘t he usa-ni l r e - a - a - wa - n k a aetscn as of c a - r a a - q c s t lon at-ac conse a-aame -s- ‘ be p e r f o r m a n c e loss a s a y e-X Ua - ct ’oe cia-a- a- - a r to a- n a - e -r- s Suma c c o n t r ol s are- added at a-and Le -t w - ’ am ‘pa t e W a y s I f t ti e - a - a - i s l e to m a - n i n a - i s o “ t-Sa- S L e t changes i n t r a f f ic r a t e w e e -n o n e i amir of crna - rraml s d am e t o interfering traffic by sarr i - amm a e r Si a-- - a - I is c-ira - ama - a -I i s r c a - p a i r e d oar a ‘ per networrd e a s i l y f a ced g a - a t e - W a y S oa- ev e n ra- - a ar m ‘w i t h per coma -arc -c t vial a- t i Ma - -re - ov a --r i n c c ’n i c i t c ’ a a to achieve re-i c a m e - I c t r arrisa iss a- - ma - i-in ” a a - ‘ ‘I’ h us ur ic is - a - I- v ca- ’ r t r o l c a l i — a s - ’e t t ’ a- I s is a rks in the a- a- nr a-a -a- - ne a- ’ssenl s- t ’ ’ t W s - a - ’ r p a f e - w a y is a na-a a- an as m to ha ve v i a t u a l c a l i f - a - c a - t a t i -a - a f o r r’I c ’w c u a - a t r a -l - e r r o r r enom’cra a -— accounting a - t n u St a t U S a n f o r a - a - t s ’f l in view of need further ‘s- - a - a - 1 2 i ra ’ ’ a - a-ce’ t ha t n5rv - th e i ’r o c e a - a u c e a - a for’ Con n e c t i o n s o f X25 r a- - f e-me s C ora r n ’ a-a-t a-er a of 4 C o n n e c t o mt oil Vi s t a m a - I cal i i n t e r c i - r a m i c e t i o n of d t t ai ra Tr 5 V e exp ect we w i l l m iuVe - - to I va r tamal ca l ta- ot i ’ ea i t t - Ic a C c at S S ‘ r t a m sa- l c a - L I mae - a - waS-a s r a - a - c ma - —’ a’ aa ’ m - - e- n a-c s n-a e a c a o t a r - 5 n L -t w t c r as ‘ v it a i t a g re e n e a t ’ - ’ I at a - e - t - S t ’- i a - a - ra- -l a a ma - a u a y i m t t e e- sorh a-n-i ar e a - s X25 3 the - foli- ’w t n c 5 aa - c t ’ c orh s to e a c h a - n a - c r c o r a n u - c t a t a - a n s of via -a r r o u i f a ia - ’ r s o f l i a-n 1 m rs ta-ata - a-S IC t a - as - a - a u to se-v e r a -a s a- c ree -t ’w -ee a a r t c m c - a r a r a - c c e a -’- the rr a- m ac a l ’ rj es d i f f i e u t p o l i t i c a l l y ant a - - s a - a - ta a - a a- e- c’t ‘ 0 0 1 1 0 ooe a- e - s r c i m t t a na- - i n - i d e r t a - a te - of f o r t i l a t u t mSi S ir ea a- i L w u t cn t h e a-a c t - a v i t -s o f o t r r e a r s aOa- t t’ caa a a a - l t r o t a ic aa -amte a - i i ’ s- t oe p r o n e-I Ln in t e r c o n n a c t in ’ “ i’ aib l n c ’ We expec t to cor ’ c c n r rt r a t e - er d of 5 ‘wa t a r t- t m a aa - co n n e c t io n is na - - ra t e c m n a m u e - s procedures ir’ ’ I appe a r to t h e - Wa Li taunt to r ary - -j —— _— _ mo ve in ___ _ ‘o r s-a-t “ i’r ’ - - a t a - - ” X s3 r ’ c s-a---’a- ’r- s - mr aa - m v n r a - r t re - - n s - ‘0 by v n r t n a - -n a i s e m a- a - - W a - at - a - as X2 ii a a-ia- a r 1a - - 3 c o n a - f a m a - a a - e ’a- ’ t l e a m a h c iy a- ae acne a at ta --re - ‘ 5 t a t -on la-ass ‘ t ’ t IO a- i a-c ’- -work w i t h c e t s a r i -‘ c-a-Il cr reasons call r e t 5 - -ric s -- ‘ _ a - i - —- - - - ’ ar t a ’ UaO ’i t - La -C a C a- aa c-n ’ ’m a - e s L a - n -s ar a -n tha - ha - a s be- a - ia- r a t t l e - i 0a- a_ a ’a-u l i s m t i a a - us - t a - i s f a ” a - a - t a - a or ’a a - ’ c a - t r a - t t r a - -a -ic-a k Laa e - ‘ a-’r - is r - - - u a - -a- a s e-i l e - l c a - t a - are a -- ’ ot ’ t ahe is oar — _—- —‘-— — - --- -- -—- -‘-—--—--—--‘ - — -‘—-- tn tIa --a - a- - ms “h af f or n a --a- a r _c u a ’ t - c-a - ‘- -a- ae -’ t e- ’-ei aa a-is Laa a- t r a-Th ni a ll 4 am n ’ c -n t - S b - a -ma - a - Car - - a t -n - I 7 6 inc a -ha -- rai a- e s a a Ira -a-a-a m en V i r a ’ s - - - -Ut m t - -a h a - t i r ‘S i l t t i a a - t i c a l l o ’’’a-’I ‘ c - a -’ r I a n ’ a-a- - a v a- s c a - a I am l I at a-’ d c m ’ n r a - a-a a- I a -ta I I I i e m - - c - t at ’ a a i t r a ’ a’ - a - ra a - c - a - a c a t c a - a rra - -c l i ’ tfl ta a- ‘a ‘a a usa 3 a a ia am Si ’ i” L - ai ‘ e ta - - a - a - in - a l c- Ou t - t o g at rh t on e a- in -c i s a rv- -ma a -- a -a C a m’ a- - n- a has e - i a - 1f l a a- t 0- f l ai r Ier ‘ idea -u tar s e - I L Sl i ’ a - at a -- a -are a- ala-’ - a i n a- - _ t or - r a i ’ aS _ i - li I 7e- a -- re ‘ - t - ‘ a- a a- ’t 15 7 a- ie - t k ‘u i f l a c i - a-r a ia - a a ’ S a is t 1 -a l— a- — 1 la J lia- o a - n-nsa - Se-a l t a- a- ai i m a- w s i ’ a- 1 a -al of j s i m ’- “- its e-’ca -c ’ at V e - l r-s - it -ma ’ ’ an t a t lIST - a r-a- i t s I w u r a - ira - t t ’ i t i t - - E ia ’ I a- ’a - e I a n to 5 1 ’ t - a ’ t at tS2 a I a a’ ’ F d a L e - on e ach SI _ Li a _ -a a - ma ’ ta - a ”T’ - - aI -t t a t e - ns a - a - Va- a I - t a - c a-i t’ _ _ - t ra - a - t i a - - EI ’SS 1 n t as c r t ’ ’a -- aI c a l l s a 0- t asS el ‘o r a - t n ’ -_a l I i ‘ n a- la - a t a-S Ccli i - ’ rt n u l - ’ -t is of t- f t ir a- a i a a-i t - c r - mia - e t a- -ac in i on s r - ’ ’r a a- - - - r on 1 e-’ ’ t I ta - t i- - i asia - a a - niIna a-rI I I as ’ t ic ” JVO S ira - n t - a I ra - s- a - is t o t - ‘ ata - - ” - r il - a - -a I_ i a e m - ‘i t ‘ tat tar 1 - i n — fsmken n s _ II’ C L ’i’ ’ t ’ d i t i re s o e I l A f n - ’t’ _ - d - C ’AO l’i i’ a ’ Pd a - ra -’ ‘i t i l ’ h l i a - a - m l c-a - - n t eta -V ia - 7 fe - sir - i Ian pm i ’ a f it Of t a - h i p1 a- n I f - a _ -- I ‘ ‘ a a -I t v i a ’t r u e - I for as ‘a-n a-h -It tu p a-nc 5 - as t -‘ - r a e t-I ’ t C C 1 l a 3 ta-a- r c cr sa - a - r c a b ’ ie-r t-a at Ch am ‘- - v ah i ’ c a - a r f o ” n- c h c ’ t — isa - i t ’Ssc - a- i e S t - r i1 a- i r a te - F— an a’o I mi re e-- Ira or c r e a - g ae re-a - a- i nsa i - n’ a - d i ct’ f o ’a- Ica C- - - a a - Irm a-’ In - cc ’ - ‘its i a - - i v r t a a -i ‘ sa t Ph c om a -nt - - a t ion g ir t i1 env ’L n a- ’ a ita - lI sa- f r o n a L — e - n i r n l a-m a- 5 a - lf ’ -m nt -a t - r r cj a - b a - k t r e e - f e ’ ’ sei - -i c-oa -- c a - f- a -tic - era - -ut I a se- 1a - i 7 ’ - tme c h - a n i s m a f i r sa t ca - a t a o s i t s e - ia -a- v 1’ cal i n ’ a ” u a a - t a - ’ a 4 2 1 fr t at ’ C m l ha - v e ‘l c n a - - ta - a - I a-a-’ ’ - a ia - — a-O - 4i t ’ I i t c-a ’ 4 - 1’- c S a - n - -a ia - a l t e n a - ’ i a a f i r aa - a 4- ’ n ’ - c t - a ’oI a a - aI m a a am -a qa - in a-a e- ’ it a s t e r - i a - a t t an t h -t n th is a-a ar ovi Iccs a-c m i-ar’ - - I’P I U se a-m a ’ a l s s i o r a - u - a Lr icn a- - sa i i ’ - a-n t h e a t e ” a- - a - a - t ta - t l U aV €n i a - a - ic ao ’ r i t a - - l e v el a t ‘ a- e a - I n a--a--ct sc caa a - rn I i t a-a-al f a-h mae a i i r a -’c t ’ a--al -a ’ L a na - — -- — -m r— f o r — ‘V e - l Ot ‘ -- i n t e r a C t L ’ e aa - a-a I_ n a- r c It ‘ a- S t a - ti - SI - m na _ a l - t c m ra -sj - ar Ia a Ma -l Ea In r a - a a a - i -a ma - a- a-i - - a - C i I a-at - ‘ m - a-a-a- t m a a _r a -it a-i a- 5 a- r a- _ a - - c m a t t5r I int o x25 a-a- a nd u s u a - a i i y lev e l ar as a - a -mr- hut I ’ S-a - ic _ ‘ - r a - a s I a - a m a - a - a - t ia C sanu —a _ - c — era - cl m - r- a- t - c c e - i s access i n t e r f ac e - ’-virtual cr apl a a - a t i ra a- a- a a - a m a- a- a c t a-25 laK e ma - a - a r th a-- f 0 1 -- 4 ac-I l a-’ a - e - vcl s- s t a - C a n o a - t c a -a l l a -li _ e ia t a a - i or oa _ as- n anraca -a O fl c a- ra-a-oa--e- ta- - ca-all f a - c a - l i t s - ’ or rr a t t n a a - a 5 is CI a a mh a- - n ’ a- s i t u a t i a - aa ‘ r a e- ia-’ cii I a- a i a - n s aibl - - 2 2 ta t a so f t w r - 1n r a - t - r - a c a - L r va - I I IC a h aVe r e v i sed a n d ________ — - p a - a I a - a -- m a i a i a S cSi iI ‘_ ‘— _ — -a s o ft - a r e h a-b l iP na_ I - ‘a -t a - a- ’ ama - - v i m a u ’I i ‘— are t a - ” a-a of hosts ca d—t a - - a- ea-- p a a - - a - a -’ a - a - - l s a- a- - sa-SS a-aa- c a-a ‘a --ct to ta - ca - cia of t Ia-ama- a ma -a - Ii ’ i- a- - — h a - i a-N lI t he t a -ar a n a l - a - f a I s ’ a- mc s ta - c c t u r e m a - t h e G-cL a- - w e -a- a_ ic aas -a e a a - Iaa - ta -a-- a - - a - SO Vasi a-a 15 r i p i ra - a- a- I a m ed a - i ra - ‘ - m r l a e -m’ itt 0 5 - ac I t ’a- ti I -i l ’ K a r - li ’ a n 19 7 t h ‘a- a s I a - t a - i s s e c t i or m a-ct I Ca-’ cr ’ isa - ca - a a - ’ a- ’C U l i s - U s a - r ta -- a - s t s ama-as s- - a t a- - -a- a ’ j I a a- ’c’aa - I t t a - ‘‘e-a- t s pr e r e a - u i s i t a - 4 1 -a-- a rc a- p an- ’ a t ia-is l ’a-s-’ s i t a i - s a- -t’ — cc- ’a-l a - - te a- a t a - a - i t i e V c a- s n o s a r a - a n i e a- a-e a - f e - r e n t s a - n a - V i c e la - - c - - i s Ju r i n i the- ne -x t cta -um - la - ’ of for a- n _ t i e- uf a - i ’ca - t — a - -c r - -a -- - i r a-ci i r L a - a a i i a - t r e u u - t i o f l of c a- V a - I S i t a ’ a a-- - -e ’ r im a e - r a -t n a caa- ’ a-- a- -_ st a- ’ a - ala - a - s r’a ’ a - at v e I ’a- ’ - i 11 I - r e m V i e ’s Va -I CV I a - a - m n t a - - v c a -’m e a c c a loa- 1 -a Inc-a-c t I b i a - ot ni rocac cum s e” ’ - ’ -v a - ’-r a-aa - r ’ c l c - t e - i a t a - a f l C I a - l r C i l S a - lt owea i a - a - Vt ’ a - 10 t a - a C as n a - c o - s - - r c - a r m at or - ‘i ia - a t ta-c it a-c’ ic- a- a- ra- s c a - ’ s - c e -s e s se -r a t e I t ‘ i a - D c l a - l Ca - fcrta- r a-a- t_ e ta sa - t - m ’aa -’a e s a- anm ” s’ -‘ n ’ - t 4-oCr - ‘- ‘ p r o t o c ol s li sa-mrs of i ’ s a-c- c a - l o s s -a c r -a- a- -s am e j ’ c ‘ - -a-a-i a - t a - a - F O n t Is L a - araa - i — i- ‘S-el Ci -a - -lea- -a- an a -S a- a - n - j a - r - a - a - a a- ii ’_ - a a a- a - a m - a - i r a a La ’S ’I a V a - ‘-‘on ‘ha-CF a- ’ t-n cr ’e y a a- I a 1 st e - a ’ a_ U t - C S a a- _ e s a r 1’ - ’ ‘h- a - m i s t a ’ u n i v - - a 5 1 t a c - CC’ a- r a W ’ iu s e e s ia-a - ‘ “ ‘ a - ae - t - - s 1imii ch a s S C c i a t r a - ’a t a-- I t e a - - na ’ 55 15 n S P - I Io- ’ a c-c uae--a r c i a a r - I c - -ra - e- -- to-- i a -n O sa - a -- r -55 p r - m r a r s r r r a- a eve - al a- -a -a - a a ’ 1m _ c a -at a-s a- ’ ac a- t c1a - I 5a- J ta-a- 15 o a- ’rr - w i S m ’ - n ta - a -a - m i t - ac-vol i a - e l _ -“ V t ’ r a - t - c - I s cn S a t a- - a- at - ta - a-it ‘ea-a ’ - a O S a --e - - ----- a a aa a i -n n a ’’ t a ’ e-’S ’ a-a- 1 ASh pr ip- -scd 1 a ‘a ’- - a- ’ta - ’ a-e-ea I” t h i s -a ra a - a -l 5O n e j a- a t 5 a - a - i5 C I ’ a - i t h e r o le s of ca-na -a a a cc is Ce- - n_ l e e ‘ t a - c_ a _-a- ’ j ‘ a-i V - o t S h a - c a - a m a -C It - - t a- a r S t a - - r t i t a - ’ cI’a - S h A D E S X 2 3 an I a a a 1 - 1 - ia- t a - aa aa- e n r -i e-X a- a-tit c a e- t h e- r e- Ia- ’ a- a r a r La-a-i S a - a - a S nsa -a - ii t - c V l e a ’ ’Ca-i - Ca -aa- apa -- c i f i c a ‘a -y i n s n am a - L ea - r - c t W e F r a - a _ a Ou t n o t recent 2 t a a a- - S - a t _ I l r a ia - a p p i n a -i a - - f a-a V a - - at r c ‘a- t - a -’ a ’rmn n r - a l s -—- - — - -— V I a- I’ t c i a - i a- Ca -i ‘a- I a - a I a - t ’ i n s o f l l a n a - n a - I C a - a - C t - i b eej’ - - ‘ a-sa-nj a - p a t a -- a a-’ - __ in S’ a- - - a- a-- “ - m ir a -n a- C aa - —l ev a - a as- tta - a - e - SbI ’i e ’ ai a i i a a - C A T E a- a_ he a n ra- -w m a-a- re -”a u e s t a t i t -a- - r -ca _ r a v e - y e - C t e - a - F a a - a -I n - - a - e s a- ar t v i r t u a l cd l i a-a- to a c e n t r a l SWITCH - a - a n a a - e - cr a-a I t a - ’ ’ a_ u i n a- ia -a -Na -iL’S a-a- u t il i t 1 w i t h iso r a t c - ‘ a - l a tts elL is i a u a - t ’c-S S i i m l f r o m ot i a -r at future is Jloc a -it tt a-a-ails on c an t s- ac - t lee a-a- a i q i n a tu r a-aCt el t aaci ’ c a l l c an be f or wa r ded Ca n_a-a F o r w a r d r e q u e s t s t a a I ra - o r d e r al so ‘l a t e - - ’ _ y t ia - c a - 197 -13 ‘Ia-lm r ’atOra - tare ide ’st lt a - of a-is La- ’- a - a t t a c i a - e - d a nd ra - c - n c s t sta g t a - c ia-a- LOCL asi at i c a- a ‘ o-’ i er ra -r c c s aa- - ‘ - t s ‘ if a-ao l a - to d i s t a t n a - s u i s h - a - a - a - ‘-c o u t a-C-il ‘a-Cl t h u s arc-i a -I s t h e t hat a-lest Inatica a-a-cu w i l l - h- a p t - u t a-ia-ta- o r r - a - i ma-a-a- t i nc cre t a- a- t h e uia - i a- ’ r Cv a- aa 1a - s -a-f Data - a P i c l au r - - i a - l i e - S b A a - r -ra at e rac-a- saaisra s e m ri st i r a - f o c i’s t h e - a - ta - f u c t i a - t i e s r e -a i a - r ’- t h - -or a- - j a - a r L o a - s- ’t Lil a - ma- a S e C tall hos t V i r t u a l ‘I’ er ’ n i n a i t a - i e ac h a-ire ta -a - el to ‘a -i at’ l’l’Cli is - S e i t h e r call is unsuccessful t o -I a--a- - n a ’ - a- l ’a- ‘a-he - s c i r a ’c - n t l v S i m a a-a-a- a’ a-a of two virtual c a - a l l i s SUC a-a 1 a - a ca - f l t i t y n at ad r e - sis a -h e i n f la- i t t e - a-a ca-a a - I a -a -ut to that a-a- ’S t l i i a - a t_ soft c ’a a ia’ a-- r e s a s i a - r ’st a - l a - a on t i r e c j a - I t e ’a c i a-a- a _ a c - c a - l i t li rca-hus - a - n _ to SWiTCH a-t a- f h a- a-jest anatictn among Iaa- s t - a-- C- - r - - - -a t ca - n’a - e - f l t i O t i t a O a S t i f l a - - a - l S i r of lox Ossa - co - a - n ama - i l in -- r a-a-guc st a IL calls a r e A c a l l made t o t h e g a t e way v i i i r e s u l t ire -an n_ Isa - r e c i p i e n t a- ateci a-n a - a - t a i l - _a- —-- a-- _ -- A SCI I r a e -tw a-k ra - - s st r P ru t - a - a - a --- l s a t l a-m l e -r en n - a-i ire t ie -n_c a- t a r e t IIflO -It a -3 5 i a - c a - -a r r i v e a t i II’a -CIi w i t s ’ no I a r c - t ac - I characters a- a c m a t r i c a - n t a in e a i for a nd icat mna - a- ll c a e st i ’nr a t i a -ra a ma -a-i e a r - o r inr for i ition a-a re a- also A SC I I st rm n- a -s S a n t -c t I - ’ - t d ma - a - as c a -n a-a- - - a- Isa- cu r n _ r-am - l c n a f o ’ a- s a-tr t a a -ma a -n d i n te r p i ’ a- -t aed e - r aa- - i n - a t e - _i - in h a-eyed wit bo ut b 1’ a h uma -’ira- ns - ’-r a end—to—end cu m ’mvt - m a t i o n a- t r a y s o f t wa s a -c a- at t ir e ia - a -Va - i 055 t h e a - I ir ep r ia t e commands cai n Ia- ’ sup - I aed ia- la‘ ib i s ch o i c e is a- a -ut f u n d a m e n ta l to t h e imp ia - -m t’ m ’ at a I a- i ore C -u t t h e c o mm i s s a - o n in g e - a - t SCS a n-l from A basic user the cr e a t i’ of conn c a- ion a - d d r a -- rm sin -’ -a- a - - I error codes could be devised to p-a-ss or i n i t i a t i n g s o f t w a r e to t ia-e c ialaa a- - w a - a y of codes a-san n t - r c o a a - a a - c - c t ia - m a a s a- a il’ s s at tss fort -a a -ition r a - e c e s su r y b e y o n d of c o u r s e be e s s e n a -t 1 1 1 ifl a-i p u r e c a ll H- a - i n ’ a tj at ’-w ay such a s at t o q e t a- r w i t h s u i t a a - i m ha- a - I j a ’- j - SC it l a - i S to a l s O L - t f um s - i a a it - r a - I t l n a - e c c s s i t y a - ’y uha ae conr ia--ct tun a- f t 1 1 w i l l i n t h ’ i nt e r c o n n e c t i o n of X 5 rai a -Lw o i la du c t -a a ra - a t is - s a id ‘ I’ a — j u r c r -iapj-in - - I slice -isa t h e ’ PCi - softwa -trc l a -t a -t w e e n l I 1 a d essentiall y a se t of to keep La- - -- a- - a a-in C AR PA NE t ’ ra a- ipp ing n i- - a - i - _ a l e - ‘ir a ’ a a -o ’ a- to pr e ’v i ha - - - c O n t r a i n_-a- L ia - Ci se -oct r -ara - j s table b ut i s a- iscl la- a - I la-a- t -r e ntj a-tro ua- check on orderly es L a - a - I is-ha -m ont a m i d - _ _ -- a - a- n is - 1 -- a-a interaarice a - -_ _ a - --- -_ a- _ - _ - I NC 1 ARPANL f V R’ L SOFTtJAPE iAa-P PANIs-i’ _ _ T I s ’ ’rpcj J _ ‘ j’r-a-5 t P S S _______________________________________________ PACKL T pa-p F P’-tlX a- SS a- ISa-Il a-h- a-Cl a-la-CL L RECEIVE 5 DRiVER Fig 4 1 t L Softwa ire x s a 5 a - t - c a-sea-i - - - to Provide a a-taon i na-n e-’ - t wta- -n EPSS and - P- a-i a - — 3 0 -t n l L - naa uma - p e d call of t aei’ st a - i t c u ca - a - TO tla a S e n d SWi ’ l iE i a-na -sa -a- s-a t e - s a nura - L-a -- r w h i c h t h e ma-i 1 saa - - d ca-e f le ct i on car ’s ba -a a-a- a u ra jf -a - m a - -c t ei to ca-c ALLOCATE OPEN a-n o r i g i n a t i n g c a l l i a - a s — i de -r at i f i e d i t s e l f — an a-attempt a -it oa- a - a - aai conna -’a-- -ct ia - ca- n ii a a-i a- c ING i_ s OPEN — t h e c a l l mapp i ng CLOSING - a u s e r or i g i n a t i n a - D I S a -’ONN i ’ IC” — c o a mn q the c a ll e - s t a - i l - i is b c i aa -g a sh c d si g r a-al hat s r e q u e s t e d t h e uf t i r e c a l l a s i na- c o n t r i b u t in g i s a a - - - - - a - -nr b u t Ca - aCm ’ s ca-a l m u s t C’ - s aa a - a t a -a o’is aa a-a-a t a i a-i t e l v ‘ta- c- f i n a l pl si se a- a-- a no n or rr a -t l c a l l i n t l a - e - t”a e - t W - a -t - i w h e r e t h e a virtu e a is analogous sina -p lex p a t h to a wa-a 1 1— k n o w a r in it ir i l cc i i a-f r - a- a- aa- a-a a - n _ is a- dJ ia -CaSS in t h e d o s t i n a t i u n n e s t t h a t h o s t la -a-a s a- a l l o c a t e d a su i t a a- l e un i v when e st a - a L i i r a - r e s e r t a - n a - a - s u e a d d r e s s a-ir a-I t o t h e o r i g i n a t o r can a p u r e d-a a cx p a t h in our c o n t e x t t h e m n i t i a-at la - a - i of ‘a -are a-a c a l l c o m n au n i c at c - a - i i i h a c k a--a - S a - Itci - - p a - n - ta -a-a - e x i s t is a-in easy stepwi se- process w ith d c ’ steps to be t ak e - r i i f a- h a - a - n i t- ta-he c a l l is na- t e-a- osa- ’i L e t La- cl a ra - - -a’ a- n -ai1 t o n s c-a-i ‘l ice cl-a-a- s irr - 3 o f t h e c u l l is more comp l i c a t e - a-i w h e n om ac s iaic a - s aa - gn a - a l s t h e c a l l c La-a- Ira C ca-ciado of t h e c u l l na -a - st a-jo I r o a - a q i i ‘air I a ’ - I ’p er a-dema -t -a - ic - ca-a-nncct proce ss In SWITCH i t h a pp e n ’ - a - ia - a - ia I i a - - ne a - - m a - - a-c ’a t cal l r e c u r - i s c r o s s — l i n k e d a -re m a i n ta i n e d f a-a- r c- cn a -a s-I s-saap a-i rtg I- ’ c-r t n _ a -ca d u c a - i t i o m a - o t h e c a ll much of t h e i n f o rm a L i th i s du p l i c a t e d a ’ t t se-a-- a- uc a an a sira -- le record would suffice- at one St a -- ra -- ’ -a-e c o n s i d e r e d i m p i a a - a - - c m a - t a t i o n to a-ca more e c o n o m i c a l on core st a re-- Thi s caisconnection process h ow ever involves a-a synchronous changes in c- ach contr iCe -a -tory call which a - a te - in a-a- ca- way u s- a la - c c_I l a - e r o t h e sep -a r a t - ’ a - a - a - all r e c o r d s a-a-r e es entiai a-a fir we ha ve had a- a c- O a-a- ’ r a - a t sa -ra - a 1 e x p e r i e n ce b e t w e e n ARI ANE r I usa -ia -i the sWiT c h n i e c l a - e n i s m e t s-nsive ly for H P A N E T ARPAN aI ’ 1’ m a p p i n g As dam c r i b e d in Se-- a - ta -i a - ta 3 t h i s ut r t u a l -and EPSS h - a la - h a v a - ‘ic- r a t - i r a - i l l f i t e- r t a-a ’ entir e -a r e l e v ant s e a - Chapter 2 _ _ _ _ _ is a- a- c a ny o n i t -a a t vehicle for UCL system h a s b e e n c o n v e rt - a - a-- i _ _ f r o n t —e n d i n g ta - c d SN I T C H the interface a-a-a- - -“- - ‘ a - - a - — — -———————————— One oaf tare a - m a j o r pi’ a-a-a- ——--——-—- —_ ---_ —a- -a- I e xp e r i e m a a - a- -a-i h a- s ac- -a-- n t a - a-aa a- Se-S tire t e r m i n a l p r o t o c o l of Aa A NIa - f is o r i en t e d to s in q l e cha raca --a-rd This protocol ces igned to m i a - a c i a - in e s s u i t NE C ln t i a - v r t e a - a - l i e - c as -i character-by-cnaracter exchanges between u s e r a - t a d h o s t subsystems one can type a character the r e m a i n d e r of t h e word t h e h o s t w i l l f i l l a --ut ar ia -i T h i s i s a vera- ma l eas unt f ’i c t j i t use b u t when in ud a-i ’a -tion ta-ha - - b a - a - a s i c mode- cf host also p ea-r a-a -ir s- a i Ir a - s-o n 1 cia- a r a c t e r e c h o i na - use tk e r a- s to ttre to let tr e e la - s ad o f the network in very short 1 tackets bucus-a-ca-a- e-XCa -a-a-S 1’a-a -a-j a- - m i ara - a -na -l n _ i a - c ’ echo delay long in our ca a - a-a-’a- t i c i S i s - ‘o a - Ce ntu a- a- c ’a - a cu r e to u se o f a communications sa-stell i te i n t r o d u c e d by t h e e x p e n s i ve l ’iTs In t h e Win _ n’s ti es is aatudta a n t c r n e t w o rk p a c k e t s c o n s u m e p a r t i c u l a r l - a- - t h e I a - e r — p a c k e t tari ffs i-c s- na-a of use is- -a- e n a v i r - n a - a -a - n t l a - i r a -j a m p a - s t a m ’ i s a - a y a-a-I Ia-r’ h a - b i t m -‘e- 1 a- - a - - s C ar ’s ch-a-rcicter r e s o u r ce s nec ess a ry to f o r w a r d cat ch p a a -u e e -t sep a -a -r ate - i y a-a -e -ar cia -- a t i n a - m a p p r o p r i a te a ck a a - ow ia - a - a - ic i cr’re m a -t s i n the p o s s i b l e need to am a - a cm e a-i t t-c -a- a- a-arc r e p - a c k a - a - a -cj- ca- a-a -as p u - a -h u t m e ot Is-c is it a- t ne vie-is of n e e - -e s s a - ar ’ i s-a- homogeneous n e t w o r k t h e p r o c e s s i n g load a-a-a-a- v Ce e x a - a e r n a - a t a n t a-a- ire f a c t ARPANET a d d e sig n a rew t e r m i n al r r o to c oi w i t h L o c a l ec h o m’s - b u t t h i s has never n een f u l l y i np ftm n’se ea nm ea -l T h e trend a- a-a- t c -r m a - i n a c o n t r o l l e r s h a n d l i n g waa -u let l i n a - a s of H i g g i r a s on 19 7 6 D ci a-ADs i n i lur o n e t t ex t m oam n s traa-t t a - i s p r o b l e m w i l l n o t recur an u -a - a - i t o the same wa-y w i t h other networks Siow control bec a- a-se me s s a g e er a - a cr e a-tcr - SS t h e a- a L a - a -e J c o n s - a - a a--a- a - i o r a - a c r e - - a - a -s t a r s - ia - a - as na-ca t h - -m a- t p ro b l e m aj a - t a - ’isa-u 1 fa r i a -t c r n a 1 a - n a - ta -a msr a - g-pa -a - a- a- se Wm1 s cit are a - t a - n t a - a - ’ as- a t ta - a -a a - l a - a-ui fe-rca-h on i lda - is - a-Ic n _ a - ex a - er a- m a-au nt wa a-ia- d i f fe r e n t aaai- u n t s of use b u f f e r i e a- a-i rm ue nope s ho r t l y on p a c k e t s in t h e no rma -a I s t o r e — a n d — f o r w a r d ca-pa -c ity ‘n_ lee process-a -r o v e r h e a d to t r a n s f e r m e s s a g e s to a-a-re d become e x c e s s i v ebut removing i t yet a d d i a - u n a - a - a-a s mama-a- a- On ca-xper a aa-rrm t ane g w i t h e n c o u n t e r ca ca - i a ’ a a - t au ra- p r o b l em s in cr e as a - a - a - i - a - f f a’r i r a g a-a s u i t a b l e f l e w control i s an e s se n t i a l c o n s ti t u e - n e t of t he g a t e w a y SW ETC H m t rj tilC need L a qate - -’ a-i - r c a-pa - ire c i a - c a - i r a - ia- rag p a c k e t s w i a -cie could ti e -l ay across t h e -pa - h c-w a ay a-rom e - b a c k i n g s t o r e ‘- i s i i l o ur re qu ire ra --a -r ets in t ia -e tate ct - fac et - a - - a nd I am m a - c t w - r C a - w c tia- 01 f f t-c ’ra - t ta-aX — -a-a- a-a- a-a-se s-ic w s - l 1 ‘ Sici ta- a - a - li in a - a a-lea-s i-a- n a-if a mum me ssaa-a-c l e n g t h s h a s na -- h y e t been e n c o u n t e r e d a t UCL a - a - na - a - i ca-in be avoia - i e a - i in an int eract -a ve t e r m i n a l m a p p i n g I n genera -al h owever fra gment- a- a- a- ’ —a- La-__ _- -‘ a- a-a-a- - ‘“ — w ij i be n -e-a- ’_rc a r J a - a a - s a -‘a k L I’ rrauiL i— p aac tie a-It m e - -s s a - ige ra- a-’t 5 being Ch a -ap t a -t - T a - a - a - a- ’e l U p U t - i ’ o a pp l i c a -a t L t a r a a - c’c - r L a i r ’a l y Cia- me-c’s c o r a - U n a -i c a-’a t tr at o a -- f a a - a - t u r a c t i u m ta - ire a- f uc i resiu I ira- ’ service a-a-nd is usa - ’a - ta - t i- - a l ta -a-c ’ L a - a - i s f o r a-li a- - r t a - ’ a C t a-i - Va - Ic tiara-a-a-er ’ a-t -v t - m s c a - - V a - a o a-a ‘i i a- i n t a -cs - a at a -n a- i- a - a - a s -a-a la- a- ‘ s’ w a r d s L e a- ‘i r S C a - t e- is a strong a-a -a-a--a-i a ’ a - s t ’ -afa - t f i l e t r a n sf e a - r s e r v i c e is t i c ’s a - a - a - ea- u a ’ m a- ia -” n or taa c oL - a - - mia - a - ’ s ira t r a’tap a- Ic file ‘l’ir er e Ii ai ’a ’ a’ a - a m ic a -t a - - a - c ’ ‘ta - aL ’ ’ a- i t aa - a t - -r a-i I c a- s a - na - a -i 1 - a-a- ta - t i a - L a- - tire - a- ire - i l f a - c r t - a a - t c a - t i u ct e rj s ti c s of th e a-i em - morc a- a-a- - m ove a-a-’isa- y L ila - at t h e Br i-a-i-a- ng In not rig id t h e TCP ot a-t a- a-a- i aa - f r a - r n a - ca nta - at a- a- h l’l’a- a - ii Fm Ic Trai n -s a-er f-Ca p a Ca- f l ’ef lt i o n s t h e ca-a-r-n see u erc a-aes 0’s a general in t e r n c - ra- ’ -oaa -a- 1nca ema - - 1— Lu— c nd protoco l ia- S c-er’s- flexible bu i i t 4 2 3 a - t a - c e s-o r an mao to l a - n _ SE g a-a-r la-H of i’PSS packet uc’undaries r mr a- te c e l s ia-i l ’ ‘ t t o l a - a- a - a - m u r ae- -J f r a g m e n t a t i o n h iv e s L u - of -j- s t ’mw a-a ’ se a-rca-- ra -an sfers aa -d juna -ca - t u a-i ra- A met c -ra - -r a-t lv e t - ‘ -m t’ — i r ’i- - aa a -aa - si r mc u se of a - a - a - L isa-- a - a s - s EPa- a- a - r a a--- s k a-a- a - a - ai’ - a- i is a”- e c c s s a a-r ’- a-a- a-—ta - C itm on lie transfer between to support any I ’ a-a- ’a i n t e r n e a - wea ’ k in g C- - L w c a e r a - t t a O s e t w o n e t w o r k s a- a We h a v e been cons ia - ia -- n ’ i ng v a- ‘s’ious ti aa - a a-e to cm c’v i d i n q -iram nt e r n e t- a--ic r a f i Ia - - scrv l a - a - c k— a -ne h a s ia - e ra- c r am ’ a- a-a c- r e t - ‘ a-i v e t a-u t a - ia - - l - cur present th urea - in g on t u e relevant m e r i t s of a l t t a - a - na - it VOS c a-t e h is a-i a-cuss ed below We h a v e c o n si d r e - d a-l a- a- era - i ii qgin s o n 197 a -L s u ppo r t i n g f i l e t r a n s f e r ar t t i r e CCL ga - at a a-a- ay ‘i’ i’ae i’h a-- k f i l e imp l a a - na - -m’ at at i o n a - if a-a- d irect a-it a - a- a pt - a t paa- ra-a i d m nt- - m a - a t r a n s f e r p r o t o co l s on Tenex i-a-b a-I ’d i — t a - p a -- a-a a - a - c -tat r a l i a-ca d - ‘PP ce-nt i ci 1- r ca - ip a - ibi ’- of of t ir e fil e some l a - i t t - a - St de n t a n - a t i - a r ‘l’ ba a - ’ pa ‘t a -aL a -- a -I ’SS L a - of a - i It ’- onLn a-p’ h is i t Ie ra car - s t - - m a ’ a- f f - - a - w a i l a - na - i t tia ’ a a - ra ’ - i c i o r k f or n a-’ l a - c a i y ‘ a - t a - a b l i siaa a-cr r t a- a-nd p r o g r a - ’a-- s a- a-a-n of rt i t b - ‘ at i - r - F a -ibm a nit t ime ha -a t a - a - il — pi n g of the the f i l e t r a n s f e r ia - a - a ir a q t r a - i n s f e m - a - a atc -w a - i v and at a - lie t a o s t on t h e i r’ a r a - - f e r a ’ot oa- ’ol s e e k s to a v n a d a- era - a -a ’ a- -f t h e 1- kC a i l ’ ’ - r - - t e o c - i l s a a - t n o a - a - i a - no a-as-c a- a-c ca- ’a- ’-a - -ricnc e has a- ‘ a -j a- - a-Cia - a- ms - a - - a- a- a - u v a - a -’ - a - I I a la - a - a - a - t - c a - a - ia-i - a- ’t ’ - ‘ ia - a - ‘ I i ’a i I t a -- a a’ a ’ a a- ‘a- a- - ce L j l s - a - i f- a - -a -- a - a - a - a - a -a-l a a- ” - ’ t C a - a- a - a a - a - St - at a- a- a-u 1 ta - ta- I t ’ -a-k-cr a- ‘ aa - ’ f ’ - t - m a- - t ’ Cn_ L a _ s - - ‘a - an a-i a - a- ’ ‘ ‘i - - a - a‘ r c - i a - J I a - a - a -aa- r-c a-a ’ _ r i c a- is - a-I -- u il m i- a- a - a a a ‘ at a-- a- a a-a- La- a-i a- r Ica -’ a a - a - a -I - ia - a -- I a- ana a - ’ a- a- a- ’ - - r- - - a - sac a - n a - -- ‘- rtl ma-- a a - ’ a- -sa a a a a - - - ’ ’a - r a a- l l “ a - a ’ a - a- a- a - a- a a- a’ t l a a a- m - ‘ci - ‘ a - a- a - r L L i ia-a- L u w u a a- a - s a a - - ‘ a- I a- a - a - aa-- Ut a-f a - i ’ -n - l a - - i l a - i a - c r t a - t a -a- a-a - a - a - a- - ’ a - a-Ia- a - a - a - a - n a- is a - - a - a-a- a- -a-al - t a- ‘ a- - ta-ce -a -a -a - a - a - a - r t - ”r a-LI a’ j - e’ -a-- s - a a-la -- a-a-a-- ‘ a-i a- ‘ a- me — a’ ‘ ‘ a- a- -- s - a- pt-- U a- to ‘ a- ’ b c -a - ta - a -a- a a- I a- a- a a ‘ 1 m a - n I iaa - a-s a l a - t a-- n 1’ a- a- 1 s-a - a m a a-a- a — - S ’a ’ I n _a - - a-a-S ma ’ ’ ia-a- a- ’a - S 1 a- na - a It - n a i a- a- ‘ - ‘ -c ta-I s a-- 1 - a- ’ a- - a’ —a- a-a-l I a t a - ’r ’s s - -a -ya -a-ar ‘ a- - a a- a- ta-a-C a -a-a- S a’ - - a- a - la- a -a- a - - t o ma - i rma a- a - a - i a- a - ia - a -- n- a-l a- - — ’ m ’ sit J I -go a - -- a - a - S i t ’a ’ a - - a ma a- a- a-a - a _-a - -a - a an c t a - a - im n _ e n a- a - a - i t a - a - a t r a ’ a-a-a- ’t a - i -al a - a- - a a- a-a-a- - - C -a- a t r ’-a - n r l e a - a- r ma - a- I - in am a - a - a - a’ t a ta - ma - I - a -a-ia - I ’ a - a - ’ — a-S n i a - i C - a of ie -’ -C _ l ’ i a- ’ a- a- a - ’ a - n _ I a - la - es a - ia -a- ’ an_ c 5- a-a- a -a La - a ’ a- um ’ S i a - tm ti -i -a- ’ -- ‘ a - a - t a -a- a - t a-a- t t - a - a - n- ar- h ’ t a - a -t i r a a - a - t i a j a’a “a - ’ L L t ’ I - a - - - ‘a - a - I - a- ’ f i ie m ias- a- t a-t a-eli a a - a a - it a - i a - a -- - a l a- a-a i a a - u - na - a - l t i t a - O5a - ta - a - t a - ‘ - O L a - a - t u a - s a- a a - a - ’ a - a - at ’ - a-- a i a - a - ’ i lI a - L a It a - - a -- a - a I- S a-a- a - i l a a - t a ’- a - - a - a ’ s a - i a- - l i a -a - a - a - a - S ‘ a- a - a i a t a - a - a-a-- s - a - i t - a - ‘a--’ ’’ a - a - a - t wa-a-a- - t i a- ’ i a - t a ’ ’ p 1 -a - a - aa- - a a a-a- - t ’- ’ a - a i’ - a a - I - a - a I a a- - a - a -t et er a - a - - - a a- a i a- U t am a- - a- a_ a-i a-i ’- n_ ct a - i ‘‘s aa - a - a e a a - a - - tJl ia a - a i’ t ’ a - a a-a- a- ’ i a- a- t a - a-’ - a - - s - a - ma - a a ma a a- a aa a - a - m a a - a - sm - u -as- - a’ I a- a t m - -’ - a- a -a - ta - - - ra- lb a - a- a rt - s a ’ ‘ ma- a ‘a-t a- ’ t ’ - ’ a - a -m-’ ira - a S— -a- a - a - a ‘a- - n - a - a -’ 51 ‘ r a - a a - a - 1a -a-a -a- ‘ a- - -a- a a- a a m a-a -a 1 1 a- a- a- - - a - --- ‘ a- ‘a- a-a S ara -I t a - e a - a ’- — t a - l” a- ’i ’ Ca-a - ’ ‘ fl O a - a ’ ’ ‘‘ 5 a -a- - t a 1 ta - a -’ a-c a- - -j - r - -a- f ia-a - a - I a- ‘r a- ’ 1 - - a - a c’ a- a- - a--u - genoa -Ct a a- a - a t a ’ ’ I ’ - l_ -a-a- a- ’ a--a - a a a-aa -a- a-a- i a n - l a - - a c- a c t a t ‘a--n a--c - ci “ a - a - a - a - c ’ C u a a-a a a- -ia -- a’s -a-a- r a - i a - tl i a a - t a a - a- a - a a- - t a- a - - mr - ia - - i -ma a s - - a - a -ar a-t a- a- ’a- ‘ a - r ’ - -a a-’-aa- Ir a- a a a- a- a - 2rs - a- - a -a - — I a - a m a -’ a- - a-- a- ’ i a - a - m ba - -a - a--i-- a - a - t i ‘a - ‘ a- f m a- a i - a - r a -a- a - a - a-’ 5 i t a - a - -a-t a- a- a t a a ‘a- ’ a-a-ant is-a-sit a- - m a - a - ’a ‘a-a- a - a- s- a- ’ - l ’”’ am t a - v a 1 a a- a- a ía -O r - m a- a- a-ia a Va -at -i t a - a - a a- ’ t t- a- a- a- a - a - a r a-a -I t a - a a -t a - na - s-a ’ a- a a- ka - am a - -n ’ s- a -a- ira a- oa t s ‘a- ’a-’ a it a - a-a - 1 cm I 1 - - t - ia -- a - a - i l ’ -a -- a - i a ‘ill a -c i I a -a -i a-a -e ’n’t fl a- I a - a - a - e a - a-- a s - a - a-a- -m - - a - - rag ai t l i s a - a - a - u - -a-a-a-- a- a - i ‘ a a - r- a- a- ia- ’ a a - a - a’u a - i 1 a- t i-a- fl a -i t s a - a - a - a - a - ka’- - a- C- - - - a- a- - _ a- i a - a - a - - a- ‘a - t l ’ a- u s- si 1 - I c a - ’ t l a- a- - a - a n_mat a i ’s su n a - f ’a- ’ a - t a- c - n 1 a-’ia-a-t ce - a- i l e d La- a - a a- - r 5 t a- a-tv a - - c’ ‘ a - L a - t’ a-a- a - a - e r a - t a - a - us’- s- l a ia t’ t ’ l c ‘ c c - Is - t i t I ’ an a -t a-a- i S a- a-m a’ a -a a-’ a - a a - - a-s -a - I - a-’ a- a- a - i S a - ’ a- ’t ‘ ‘a’a - ’ i ‘b aa - - at ’ - a a t _ a a-a- c - - t a -t a- a - il — a-- ma -’ a - a - a- a- “a-- a t c-a -a --a-I a ’ t a cam s - is - a a-ia-a- a - sa-b u a-a- b t a - a - C’ l- a- a- a-- a- a ra a- 1 of a-nc ’ hista- r ’a- - a a- n a -c -c t W i sea -i ‘ ‘ ‘ t a - p 11 h -a - ‘ a-a- - a -S ‘ rt- - a - s - i ar l Ira - ’ a - med t ’J l a-I’ - t a- —sn - i ll - a in -i Oi - ‘a-l ma a--m a- - a-a- a- a- a a-t - t- ‘ ii a o aa - ’r e f l a- ta le da - a - cm use- u ’a -a -a - in a-j 1i 7 7 a- i n ’ a - a- rmi - I I- - a- s a L - I - i i a - i a l’ taa - ta is as - c I a a - te l na-t a- i p 1_ a - a c e it l’ ‘a - a - a a - h I r i a - m v a - ’n t a- iU’ l a - O S a - ’i I i a - j a -’ i’a - g i t e t’ ia - u a - r i L c - ’ a t - - neSt ala-- t a- a- t n a - - a- a a-ta-- II I a-ha’ a -a -a -S Ii ’ a am I a- ’ mar oi l ’ - a-i h I - - at ‘ ‘ a- 1 a- a- 1 a- a- Sit a - a- I lets 55 to i ’a- a - a - i a - a - I i a - lI a-aa- ’ I - ‘a -s a- a- c as’ r e - a- - -i a -a - et a- - F a- rs t a b a -a- a- - a - a ’ 1 ’ -a 7 l a - s- a t a-a- a- is - a Ca--a- a-a-- i a-a-t -n c- -am e r a - ‘ ‘ a - -a-- t a-a - n t r a- m’a - a f er a-a -t -na - il f i l e tat a - - - JS a - Im ” - a - t -a- - 1 a -a-I ma - a - U L a - t e - a- ’ t a a -” l a - ’a- f ’ ’i ’t C ’ 5 51 a- a- r- s-a-b n_ia - L a ’ m I a - a-1e m n i t i o n Ta-c a - a- a- a- - ‘a- act a o l a- a- -ma - a l a -a - c t a- one r ’a a - t a- I- a - a ma - ‘ -a- a a- a-I w t i a - - l a -- a - fl L a- -a- t a-a-a- ‘a-n t or I-c I ’ ’ a- - IC - cia-a t a t ime - a-s p a - a - a -f a - ai c a-a- a- - ma - a -- a-a- t- a- oba -a ’a- t n - n tact i n t - ta-he a-a- I l l in it a a-i t r a - irast- a-a -r I ’ ’- cc ra - a-t’ - ’ r k a ’ - an a - u - s et ”l’ a- ’ t- il - ’ s l a - c- 1 _i a-a - a - im ’a -t ’ t a - a-Va-a- l a - I S - a - t a - i a - a a-- r i a a - a- a t a - a-t 1 I t ’a - I l _ a - a -l’ - - i a - i a -a - i a - I c i s -- is a-a -a - ir a- a- a- m - m’ a-i - - a -t c a- - y ’ r a a r l i s ’ a-a- a f l t a - a - a - ‘ at t-ca p a-- ’st a - a a - a - i a n_ - U t l e ’m m ha- - fl-ia-a l ‘ a - i p a - i i a ‘a - e s h it’ a - a - ’ na r n ’ a - a t a i--i a - a - ’ a - a - t a e r a-na--b 1’a - - t ‘a- ’ u a - a - a a- a - a ra- n a- t a i t ir a - p f a- 1 a ” a a - a - a n I a- r a-u ‘p an a - a I - m a - ’ i a- - ’ at ia- L ’I ’ a ’il la - a-a- ” ’i f - n t i i t a - ia - -ia -- a- a ia- n bc a - I I Ma ’i- ’i j a - a-i a a-a-ij L a -”i ’ ac am a - ta -a-- - a a a - a n_ l ma- m - a - a - araac’ra - l a - a - a - a a-- - iii a l a - l a - a l SI a- la- -a-- a r a t a-sen s i m i t m t ’ t o ca- ca-cr l a a - ti a-ri f O r n t t t a - ’ s - lU a- e a- s c ’ t f l a - a n _ -ca- a--a- I a- t t i a - a- “a-a-’tia-0a- a U S a- ’a -- l a-f l-k U ra-a-a- I ’ LIr e - ‘ a-a- a-a- con ’a - u - sc ’ r s -Ia - ’ a- ’ a -a a- a - a - ’a - a - a - I S I a 1’ a- i a - S I I et ft a- e t a a- l i ce i n a - a j j a- m a - a - a - - la - a - c - i s - na - a - a - aa - a - a - - a-ti a-t S a - a - ’ aa- a-a- ra t - a - r c a - a - a - i n_ i e aa - a r - -u a-al ‘ It n _ C a - i a- it a ’ a- am s ’ a -a --a - a - li i t c a - _ a - a - i I l ’ r aa a-a ia - ia - -a-ura - a-r-a- I a - 01 n _ t a - a - a al es a- - ma- na- a - c l -a l i s a - a- ‘- f I a - a-a-s’ ‘ - a- a- rae -s cut’l a- Sa - a a - a- t h a - ’ a- a - at Ta - r I a -em a - - ar ‘- a- - a-a- - a - tm ta-a-a - t m a - t a - - a - ’ ’a a - a - S n a - t a -a- r e a- s a- l a - t a - c —sç r - a a-a- ’ -0 1a - an i ta a- r ’ - of I a - r 1a - a - ’ 5a a i S L ir ’’a -- a-au i r a - a l a -a- - s c m -a- te a-a-i S Ia - a - a - - m a - a -n ‘ ‘re t a a I i a - a - - b a- h a - a -i r c a - a - t a - a - a-a- a-a - a - aa a- - a - a - ’ u ’ t a - a - a- a - S’a-’ it i a - u i a 00 -I a-e- c-a -n a- - a - a a- li_i ‘a- a-sU st p is - p-a- i u r ’ na - ca i t am a - r a- - a - a - - - t a-l lc ’ 1 ia- a - r n t-It -- - a ’ n i s a- a ’ tO - - ta - a ’ s - - V a - Si a- 1 ia -s f - t c i l r t i a- s a - - a ’ M a - lia - I - - ‘ i ’ hm gie nc- a - st u a - s e n ‘ rca - a - ia-a-a- a - ma - I cc l a a - r a - a - a-’ t i ’ - r ’ a a - a-ls wi l l a -t a- as S t a - e s ‘ a - Ia- d e s t t a - a - t b eta - m a— I a- - ”is a-1 a-i -o r u a - m a - a - t r - ’ --- a- a - s - i a - i C a -a- a- a- ’ - a c - i I a--a- ’ a-a la- 1a - ns - a - a - d f - n a r C a - I c - L i j a-o a-c’a c i a - n - aa - a -- - a - r 1 a-a- a a s t i t t - a - a - a m a - i i l a - a - a - a- a - 5 C ’ a-’ a- a - w ma - a - - L a a - a - t ia-a 1 s - s a-’a - ta - ca - a - ra-a- s a- - aar ’ -S ’s - a-- 1 a-- l a - c ’ a - s - ama li - t a- a - a- p a - s - c - a - a s- - C - aa a i ‘is U a- I ‘L a- - a - a - t a - t a - a - ” f i t ’ i - ’ - a- a- a-’a - ta - la -C C l - -s - a - a - a - i - ’ ma -a - a a m -a to ‘a i a r ’ an a- a La-’ a - lI l a - a - a - ia - a - la - a - ’a- m -a-’aa- la-a--s- a- ia - t ’ ta- t a-a- a-- T a -ide j ars ban ’ a-i ‘a-Va - - a t ’ a - a - a - a - i a - e a a a-’ - a - a V e - t a - ia- a t _ a - s-r ma -a - d a - ’ a - i t a - a -na - icc- u i-i a- a-’ a - t ’a - a - of ri -m a - a ca - ’l a - a m e- ’ t e rn r a- i 1 a’ ‘ a- ’ - ‘ CC - na-a- t t a ra - a -‘ a - s i t e a-an d -t a- a- a -t - Uflt j U ’a - i ’ O a a - l L a - a ” t ua ma t ’ a- ta- a - - a - a- -a I - a - -l a a - - a- a - ca- a a- s - 5 a- a -a - - a a- m a - t a - et a - a - la - - a-- ta-a a - a - a s - a- a- ’ ‘ s-a-O l a- S - 1- - 1 e a - a - a - e t - a-a- a - ia - s - a-a -a- - a - t a - c a - i a l a - c a - c r a-r i n n a- I i a - a -t i -a - t lma a a- ’ a - n ’ a - i a a - a - 1 la - a- a- - - s - n a - a- aa- a - t t a - s a ’ a s - - - a - a - I mm ’ a - h a - aa -’ t- - m a a a - a - a a- ‘ t a - a - a - a - a- a- i t i t a - s ’ a- a r t -Va - I l La- - ’ a-I l’ ’ am -ma - at a- a- - a l a - a -i t ’ n _ m a - a - - it- uta h -1 3 aa - a - a- ‘a - a - a - c 1 ta - i- -a - k a-t i a- ta - a - a-a-t a- nc b eta -a - a - r e c t a - a a- - - -a- a- -s - a - — ta - u — a - ” a Caa - -a - - fa ‘ ‘a- -a - p a- - a ’ a - ia a ’n a t -a - a - - a - ma a - ia - a - t a - n _ i a - a - a l a - a - a-a - cam -l a - i a - i a -a - a - i a- ’ t I a- t a-a-- La - a - - a a - Lu a-a- -y i a - a a -t at at a- f II ta -a Ca - ’ s’ s-C a-a -inc a-n ca-a-a-am-a-re i a - a - a - -a a-i - be sc - a - ta - - a - a - k a- n t a - a s a - a - t c r f a - a c c s- s ta -a-a-it as a - t i ’a - a - ta - Ia - a ’ - a- i a - a - a - t aa-a- to ns-c C a - Ia - ia - in _ i i ’oc es s - a-r a -c - sn_ et a - aa ia - en _ er a -a -a -cc a - a - a - a- a - t i a - a a - a - S ’ e i la-la-C a-’ m a - ct - a- - a-a-a- i a -’ I a - a -t’a - t m t ‘a-Il - a- - a a - a- - me a- a- ’ a- i i t a - l a- a - a - a - a - i a a- - m r fa - eta - a - a - — e e a - i 1 - ta - a - t a - a - a’-xe - a- a- a-s-l a - a c ‘ a - a c - a - a i- a - a - a - I - ‘a- t a - a - a r t - i - I am a - I—to— i - m-a-d Protocols a - -a - ama - a - a-Ct ronq k a aa - S a - ’ a- - t j oa-a ¶ a ra - La -’r ’ a - ’ L-a-’ r a- r a-a- ‘ -a-a- a - - a am a - ’ - ’ -i tata - ’ am a-a-a-a -’- a- a - a -l j ” ’ t- ‘ ‘t ’ ia - l a - i a - -’ a a - r ‘a- ’ a- - ua It is a - a - a a - conceived as a as - ae a- i s ac ‘5 m p -I t a ’ a- - na - a - - a - a - -a - a ’ - ’a - a - - a- a a - ia -a- a- a __ a - a a-a-a-’ t a - - a - f l i a- - ’ of a - a- - ta - a - ta- ta- ma -c mb n_ a- - - - t a -a-a-a c r c - c a - n l c a -’ l’a- aa- c t h e a - a c -ma - C f ia - a- s -CS 1 s - a - a - a - a - a - a ‘ a - C U a-a-ia -a-st ‘a i as-r it a-r e ca- ’ra- L i’ - a i a-’aOil t a - n a - a - a a - na - a -Ia‘ a-a-a- a- a-a- - ma -a- a- a - a - a-’ a- ’ a - em a -a r a - i u a - a - m a - ai- a-a -a - wa -a -rb a-’ l a-a- a - ‘a- c ’S u a- S a-fl a - a - t I 3 f l S s - I n_ C ’a’ a-’ Ia - ia m i’a- a- a- a- ’ is a-a - s a-a-’ a - a - la a aC -a- ia - I a - la -’S ‘is a - a - a - a - C -a-na - l O s - a - a - ia- a - a- a - i - a - -r a t a- a- a- Y - a- - m ay i r a -- ma - a - a -ma - a a - me ‘I i’a-ia ia -a - W a - r t- Sa - a- a - a - I a- ia -i a- ’ i a - t a - a - a- ’ a - lit ’ a - c a -t ’a- ’- - r -m s- ta--s a- i a - t a - a -- -i I I I c a -a - a - - a- i s a- a- -l o t a- Ia -a- a-a- s a- ‘ a - - m a - a - a - a - c a -n _ r a - a - a c a - a - e r ca-a- a-a- the a - c r I - - - a - a - a ’- a- - a ta ru a - a - a ’ a- a- sun _ va - a- a- ’ - - t ’ i a - ’ a-l a-y a - a - a - fi t - a - a-ma -el a - i r a- -i ‘l ’ ma ’ ’ t a ma t a- a- ta-La- m ’oI sp a - ’c ira - Inq a - a- 5 i a - a - s C5 ta - a t sa - ch a - a - a - a - la - m t ’ - a - - n t am- a’ W i t h a -a-a- ta - a - a - a - a - t a - aa - a - a -f l ’ i t’ a- a- - - a y a-i - --a- - r a - a - m a -- a- - a’a - a - a - a - a - a-a-a- a- a- ’ - ’ U a- a - i t a - t a a -- -a-b a - ci l a - in s n _ a - a - a - i a - t e a - i e a - a - r t — I c o s t p r o t oc o l s a n d p r o t oc - a-aa - a- UYC ’a - s - U S a- a - t t a- ra t a- a ’ma a - a d a - a t a - ’ a - a - a - - a - ts- ‘ is t a-a-a- i’a - - ’ cr a- a-l a- ’ -‘ p r o t a a - ’t ca a - - l ‘l’C U a- ’ - - a-a t s-s- W i n _ a - a - S a - a -L S a - ’ a- a- a-a-a-a-n’a- m a- a -a - sa -a - a -m - a - - ‘ a - a - a--am ’ a- a - s - t m a - a - c a - a - aa - a - a - ta - a - t b C ’ 1 t a aa - - sa - t t-a- 5 a - - ms - S a - a - t a - - -c a -pt t tt - 5’ a-’ ’ 5a5 a - a - a - a - a a- - m l a - a - i a - i sa - a - a a-I a - a a - n n _ c t -a - r t - i s -a - a-a - i’ i - a - a - ‘ a - a - a - a- a - ’ a - t a - ’ aa- 1 i m t a - -a - n - t-a - -L - a - a - ta-a- l a - a- a a - t - i a - mm st a a - ’ p a ’ set a-- a -t n c a - a - a - t a - a - c a - a- a-a - a- l a - a - a - a - as -- c a- ia -t a -a - a - - a a - a - a a - t i e - t a - - a r s a -1 Cr a - tm -a- a ’ a- a - l a - c a- I a- a- t I - n _ a- ‘ a- - a - t i ’a - a - l a - a - a - a - a- ” a - a - ’ a -5 a - a- a ia -c a- a-a- a m c a- a - r ca -aa-I r1 — t ’a - a - - t - a - - ’ a - tia - a - a -m ’ a - a a- l a - a - a - a - I - au t i a - u- p m a_ r - r a -a -- fcr ‘ - - a - t a - ’ i a - - ’- a ’ a- a - ma - al a - a -‘ a - a - a - H’ I a - ’ a - a - a -a - a - a - is ’ a-a-a- I a- a - m ’ Ir a - a - a - i a - i a - t a - c - a - at- o a-c - I Is a - Ca - a - i a-uc s- a - a - c a - a - s - t a-a-- V a - a r l a a a - a -’ — 1 - at ia - L a t - a - a -a - p a - a - ’ a a c- - a - a - f r o m a- a-et cca -rr era t ones used ’ s- a - t a -’ a - I a- t a - a - a es a - a - a - a -s-a’’ apt ’ -a - pa -- a - m a - a- a -a - a a - a - - a- -i a - a - a a - a - a - a - a a - ’ -r n It-m a - ia - i-- a 5-” as n a- -i a -- a - a - a a- a - a - a - a-a- a-i a -ar o a - b U a - a - C i a I a- a - a-ca-c t a - e c o r r e c t s a - - a - a - a - a -’ - a a - a -a - a y - --‘ a- - - a - a - - a- - a- a-a- a- a- L a -a - e m v i r t a a - a - a i La-’t- - a a - I a- f a - a - t- a -n s- f a-a - - ‘ - a - i a- ‘ ra-a-’ pr a_ at a -- i ’ ’ C a - a i a - a - Ca t - s- L ’a- ’ a- I a- a - a - n a - a - n a - a -m nt a - ’ aa - a - r e a -- a- ’ a- a-a Ir a - a - i a - ’ 5 i a -”aa- a- t_” ‘5 0 a- r c a -a-- a - s-t a- a - a - a - l t c ’ a- a-’ a-a- ’ ’ a-5 a -_ a t a - 2 a-V a-a- y a-a -na-’ti a -’a-’a-a- - c a - a - i’j c ’ rS c- a - i a - c o f m o t h s - i n _ or rc - a- ser’er a well i f l t - ’ r i t v SA ’I ’Ni ’I’ l ’hc ama - a t o n l y t a - a - -r e hi a - i s a- era - s-a- ma ‘is a- l i of’ a- a-i I e s c a - i a- i a -a- ’ a-roch a - a - i i s s u e s of -a -t 5 - li ett- a t a - a- a - a- a-c a - I a -a t m a t int S-- a - a - - a -fl transit iota - al l e v el ‘a - Va-’ it what Au t a Ca-i rt t ham c o n n a - -t i ra - a m a - a -n a - Ii r a-- a - nt ’ a ’ - aa - a -a - t 5 5 ‘ - - a - J f u l l -i ’n a i s a a - i c - I a- na - of b a -a - a - i Ofl iwu- v -r i’ 1 ia -i L t a- in i -a - n a - ippt - a-lCtI a t be c i n f a - r m a - a - it ao n mama - a s t r o u t mn -i tm - a t ma - I b I a - a - of ‘a - - I S u a- a fl ’ J a- ’tit 1 — t a - a- ’na - a-a- p a ca-i a-a-ph I a a i i ta-v aa- a - i a - a ma a - a - i r u a p - i I f l a - i a- a - ti Kir t ein 197 M t a ’rr a ’a - mnai ii ’ s-I Wh qa a- a-’a-i a - ’ a - a - a- o some a-f in ‘I -- ‘ a - a - m u d ’ ‘isa- lt le ’ - a- ‘ a p a - - n ’ i I - - - -i om I 7t a- r t rca- t-- a- - ’ i d e a - - a -t i l t ’ - t in Ia - i - t a na - a - a m a - c ’ n - a - - -a - ’ a- ’ a u r e a ’l a - a - aa-- ’a- ’a - a ‘ ‘la- c’ m ’a-a a - i i f f c r c n ta- k and hi s a - a - a ma ca- ama - a-f ai t he i’ i i a - a - l O n a - I t s a- a- u ta- a - a ia - ’a- -l in - a mc - d a t a - a i-r a - a s - I s - t a - t a - t- - We a-I 2 5 of n_ s et that- id- a -a - t a - a- a a - f a - i t j O i a a-- ’ j -- a a - - a - a - ’s a - a- a- a-a- a- a-ire ma -ore I’ - ’ i ’c s i st -c-a-i -mc rk ima - a - fa- ’ a- s- a - i a a-n a - a - i s a - a - is s la - ar a - ’ a- ’ ‘1 ’ a- i a - a e r ’ a l a - a - a a- a - in a - a - a - h i a ’ ’ a Ya- t a - l c - a - a - i n t a - a - r ta - a - - t a- t need ila a- L I - - c - a - a - tiI O “a - a - ’ ma as - - a - a m a - fl I a-a- i n c n a-a- -a - i l l nr os- h a - a - a - a ’a - ’ a - a - ils a- ‘t O C O I S by ta- l i e a m a - a-m’ is a - s - a - ‘ut a - a -ra m a-tt t a a -rs - ‘ a- ’- ’ t a - a - a - ’a - ’ It ’ a I t ia- a- - t i a - n - a- a- a i - a -- t s--i - a-- a i a - I a - i mi j 1 _ i a - l i t of n ra - 1 _ c Ia - a - - a - as- - - net s - ia - c ’ clo se ly 1 terr r ’st r i a - I n a - a - a - i a - a - s - i -ia a - - a -a - i to a - i a - f a - r a l C i a- a Ia- wi iua - a - f a - a i lt a- -a - ta - I tt’ r o a r t a-c’ ‘t ao - - a- a- a- a - a - s c a a r ’ -a - I e at s a - a - n a - a- r - - a - S S a - ’a- a - n - t a - a - a’ - - ‘ i ’k a - m t f ta - - s-p ‘ a - a a - I a-a- -a -- C a - a - a - a r aia - ’ ta s - n a - a - a -a n--I i a - f l f l i a - a a-- a - a l ’ a - a - ’ I a - a - i a - l ea cm- --a- a- F mr ’a- La-a - - a - aa - aca - ’ - a - t a - n e I e a - c a - a - s i d e l a - a - a - i c i r c l e s a a - a - at te a it d - a- nra - am a - a- - a - - na - -sa- t - Ca - ol a-i c e n t r a - a - i l — c o n a - r o H a’U- i ra -- r ma a - a - e’ a - - pr ’ a- a-c’ i s a - ’ l y a- c i t rca - - f t a - a C ’a a -i ’ a- e rm a S ’ ’- a- - t - a a- a-i ’ S a-a -a -a a a - i a” a- - tm’ c ‘a- a - - - ippra - a - a - a - ba - a-a-s-a ma -’ ii ia - c l I a- - - ’ a - r a - a - s a -a- a- - a - a - -t a-cor e - i en a - a - r - a l i ra - a- a- ’ t - - t ’ - - - ’r ci ’ ’ m - - I l ’ tl a ’ L p a t Ca- lit ’ - a- - a - a -- - ‘ - r t ’ ’ a- a- a-a- t his m I a t - - - a r a ’ a- i na - a - a -’ I rai - - - r a - on ’ - — ’ a - a a- - a - a -m a - cc a- I ’ a - ’t - a- a-a I I S of -ma - i ‘bt ’ ’ ” i t a-q A t -a - I’ a- p t ’ ot --a ’ ol s mu s j a - n m f i a - ’a - a - e c s r a - a - n a - a a- - a - a - e r ‘a - -‘ a- a -’m ’ a-’- a - c ’ a - a - l a - a - si ‘a- a - of ta - s I a- a - -’ I n r a - ’ ma a - a - t ia - a -mi La-sIr a-i a - a - i a -i l a s-s a-a- Ia - a - - a - t i a - ‘ n a-c ’ 1a - i a- - a - i ant i t i i t ’ Ot a - a - a - t aa -- n a a- - I m ma i -i 1 - a - - e n i a - I a- - - r e v a - a - a --a- a- a - — a - a a a - a s - a - s i ii a - a - a ca -- -o a’- ’ -u ia -il ta - a - a - a- la-_ i a- ’a- ’ ’ -dlT a- ’ a-is - a - Ia a - - a-a-aa -’ a - a - a a -’ a ’ a i ama -as ft I -- ia ‘ a- a- a- - a - i l r i - ’s a- ar - c’ra-vt ’ a - i Lb e i b ul rc ” p a - i a - a - a - a-na - a - a - a - t a ’ a - lit’ a - a - I -a I ia t - ’ - r r a -- s t of be a-It t b - a- a - s - ta - I a- a - ui it- a-a- a - t a- ma - a - a l a - i a - a - St Is Ia -a a- r - ua - a - a O 1a - ll - l i n t a- - m a - a - i n c a - -a i s ba - as - r i i a- r - a- - ’ v e r m ’ia - ’t a - - u rl aa 7 a - na- a a- ia - a- se 1977 -a- f a-a - I a- a--a - a - n t h o s t a-a am ta - -i a- l a - i l in tr ia esa - a - t ‘ a-a- aa- ta - i p t a - a - n t c - s - a - CS a- u i a - a - e i a- i sa-i fr o m a- ia -h’ a c a - t i i i a- - a - 1 i _ I a - a - - d a t i in a- nc’ I- a- a- a-- ’ a-- a - a - a - j a’j- ’ t hea- a -ha - c m o s t s - n a - t a -a - ail e d 7 _ I a - t i r a - t e r n e a - - a-a rk 1 -ia - ma of r a - t a - t a’ca- i Ca- - a - p ’ i y a - t r - a - it ‘ ‘i r ef a - lIe a - a -’ a- a-n a - f ins-i t b r a - a - i ‘- -m a - r t - n - i of a-a a-a- C ’ OU f l t i a - a - e a r a ’ m r a m n i t m a - f l t w hir - i a - m ra-a a - la - -a - a - a - - a - n a - i a- I a - a - a - a - a - - i n t u - r n a- a-t ’a-i- ’a r k - a - l a - a- as - a - ta- - a - I re s- a r a-a - l a- a - i s a - s t C ‘ a- - a-i t ’ ’ ‘ _ ‘ h a - a - a - s-a S l a m - a - - a a-- a- I a - - - tO ‘a - li t-i ll - a- a v ’ a-ja-’-r k a-1- transit This 1 9 77 78 — arc aiscussa--a-a-n further a-n Wa -’a - i c t ’a- l’ ra-a- ’ a exp e r a -n a -en t a s concl ude a-a- a - a - c t a s - a - s - a is a- liffi cult t a -a- at i t th r o u g h p ut co n t r o l a-nd h i g h a-na-I protoc ami a - n i _ a - a - t -t a ci scs - a-sseu a a -’a - a - ieVa - a -’ iia- ia-a -a - I ii a- F rom t i n e serv a - c a - a - s r c a - u - isa - a - a - c a - a--- a -ma - ia - c I ’ol ut a- ’ a- t l a - c ’ a - t ia - a - ai- a- l r a - - m a - a - -a -n t a t i - - n ivitc c ta m ae a-- c in a - a - a - ave a lev el wca -c I a - I ye a- I a - r c a - p a - ‘n a-a- t - a-am ta- c cj a-a l - ‘ - a - la - i - t r ’a- - a-iSa- a-a- ’ ’ - t r a -a- ff a- c a- ‘ a-n ci ’ must be are b e na - a-a - i o a a--a-r a - a -a- a- a l ” e r c a - a-flu t a -a -c of r a -nr a’tia a- as a- en s e c u r e d e l i v e r y -a-a- s t O a -- a a t a - a -y Ia-a- nets - a - a - th om a - i i Si a -a-a- e cam t - a - Ira -a- a- not r a - - a- a-a- i a - c a - t- a a-a- a-a- nd t h a t UCL t h e s e p r o b l e m s tn a-a-t en c i— t o — e n u t o t a - l i ’a-’ cores -a-a-on h igher — sun a - na- t a -rs -ti a -iia- i a e t e r — cj a t c W a -y ma- a-i t t ra - a a a a - - ’isa- ’- rk a- - k s ba I r am na-m t of ta-s e types rc a -duc c -s a s--a - a - c a- t a -t i ‘ strong 1-i’ — t ew a -y s a-ct 1 ng sub j a-- a-at to necessary congestion At tim e p er ip a la - - ry of a-ha - - ‘la- h a-a-sc’ a - t a - i a - u s a - n o a ’ m a-t i o n a - n w a-a -ida- to validate sub—areas a - a -a - a - a t a- a - ca- s e q u en c a - ’ a - a r ou t i n g isa -- le a - a - va ’- c a - a - ma - a- ca -ma -hac- a-i a- c’ a- La a - r a - i a -- a- a - a -i l k a n a - a - ‘ a-ta-os-n l y— c o n n e c t e d ’ set ta -a- tics-- source and destination h o s ts t a s em s e l v e s path or even a - n u lti p lc na- -t s C O n n e C t t ’ a-a- ’ a-la-’L a- a- a-aa-’Velopments la ’a- a -ca - a - a- t a - a - t a - n ‘is’ 11 ama - ” t erma -raa - m 1 a - a - a - t t a- ’ a -ia -y a - ca- ti w a-a-a- t a -a - a ’ a- ’ a - m a - a - I c i a- a - a - a - a- a- al a - aj or i t a - l a - ras w i 1 l Wa- a- a-il I r a - a r t a - ’ w a - i a- S a - n a - L a a-a- a - nc i s - ‘ a- u a - c r — n e a- only I a - a - c a - cai- s t a - ru - a-c- s - va - a - a i l s -- a - a -n a r a - a - - a - a - a-- I a-a - f i b’— a- m a a - a - -e a-s r t ore and forw a-a-ra -i na - u - a - I t -n c ont r o ta- i a - a - a t 19 7 7 - a- s’aa - a -- r r a a - t w e r k i n g a - a -’ c - a - n a - a - ma - tra - a -a- c-in t - e a - m - a- aa - a- t a- a- cr a -- a - a - a - c o i s ‘a- t - a -r o t a - a - a - a - rca - S and La- ‘ t a -i t i a a - a aUs - i Pt ia- - i -a - 77 7a a - a - ’ o r ap a - c t c - c a-l l a a- a a a- a- a- s s - a t i r - sca - a - ’ a - - t - t a- a - i s - a - - taa - m i i c’ it a - a - c c o u n t m a - a - a - i Ia -er a - a - a - a - n n i n q a - a - f l S i r a - a - a - s t a a - aS l 1_ va - a- 1 i a - a - w C a - a - f i t a- a - I a-’j a- it C’w a - i y s t a - s - c - s c d ena -ands a a - ta - a - t Pa -’t r t i c u l a - r l y a - a - t e n t b O a - a - a-ta -St La- a-Ia- a - a ma - ui standards and a - ’a- l’ a - ’ a - a a - o p a - a-lc a - c y ’a-_ m a - c a - mms - a- for q u a - a - h a-at a - a ’ a- - t - a - l l area - n network a-n t e r c o n n e ct a o n a - a - a-ta - na a -as - a -re are a-a-any ’ a- I m s oma- a-” nd—t o a-a- a-u ci a - a - s - -a- ’ t i i a - mL a- a - i a - a - a - a -a - a - a - a - aa-i- l ’s pa-m t -- l a S - aa - a c i s p a - c c- a - a limited app l i c a - i n i l i t y t yp e s of of boa-a-a- era - a - i— to— ca - ed fla- ’-a-a- t a - a - me’ ia - a -’ ama - a-a-a-a- a a - - a - - c t a c -ma - - - t na-c ta -ha -m t a--s-c C a - a - a - t a - t e r I- -rc a - r ’a-a-’c p i ’ a - a - a ’- a - a - a -a- as - a - cl ear l y f eel Ou t i sa-a- - a -’ r r a - ama - a i a U CL A j - a - ia - a - a’a-- a-s a-a- i a - a - t a - a - a ma s-a-- 4 4 -a - a -a t ia-ac-C faa-a- - t a - i ’ -— t n t a- la- t ew a - i y — G a - i t c a -wa - m y a- a- a-na - t a - a - a - a- a- control 4 - ta -i important unrest a - v a -a -cl i a - S U C S ir a of a- — TC a - i ta-c w it h a a-i single network The latter we de fine as ‘weakly— a - a - i c y na a - my is -a -ave SpL ’CLf iC int er -a - ja - itcw s -n yprotocols qateways they do not Performance and throughput on connections to ta - t ea - c ‘ weakly—connec ted ’ ne ts ni a -a-iy be lower than J a - a - Cy in the ‘strongly—connected ’ case “ a - This may still be the optimum F a-a- a-—a- ion s -a- L- r a- u r a - 1 a- -a-a- i a s - - a - n ta - a - r --a-nm a-ect a-ion c- a m en a tr-tr a- i n c _a-__ _a-a-’V ‘— V lu - ta - a - -a - us c- a - a - ma - - a - a - t a - a a - ’ a - ta- i a- a- a - a a - e r 5 is i a- a - -a- — a-——— a - sa - a - ’a I a - t n a - a --- m -m a - a - I L t t - a - a - s i t i c r a - m St i a-ha- ’ ca - - a - ma - t a - a- -art a-n a- ’ a-i t - ’ a- ’ a-’ ta- - - a v -I t ’ a- C’a- a m a - l a - p r a - v a- m a- i - a- - - a-a M a-i 11 a-ca- u t a-t ai a- a- -r Wa-c f l t - t a- ’ - i ’ a -a f ur e a - p a - r i m a a- ca-’ r a - t i r a-a- u ’ r a- - aa- a - n n c d a- a- s - cit _ a- - a- a - a - - rs - a ’m- - n t t h c - in r a- s u l t t a - IS a - í a -n t a- a-ui a- a- 1 1 - a - ca -- rk U 1 a-mt a-e a- 1 mm-- f-u n a- - r i a - r I- a a- a-a- t a a- a - fi t ‘1 a- a - a -1 ’ at a - m m - c - i so c-a I a a-a L a - a ’- c aI I - a-fl ir a - a- a- - n a - -a - a- ’ a’ a- 2 h ’ _ a f l - t m ’ d- a-”-r a- a ’ ‘aa a - a - a - i t - a - a -a - a-”t - c a -a-a-- t i m ’ s - we ia - a - ave a-ca-ca-s a- a- fl i a - a - a - a - s j t a -n a - a - a - na -i o n a- nose a - h a - t n -ca -c tints a - f lC a- ‘a - c a - r c a - -m a - a ui - a- - x p e r i a a a - ’ m - m ’ i l i v I- i 1 t a - ’ t a - a - i t i a - ’e a- ’ I i i i a - ’ Va-77 ’ta - t i a-a a - t I ’ a- a- a- ’ i ’ ’ a--au’ t n t s n m”a- p i u ’ ” - l v a - 1 mt ca - a n t - a-a- - r e s t a- i a- a - l e a- a- a- a - n a - i - a - a a- i- i i ma - a - a -i ‘ ‘ -a- ’ a - al a - ia - L a - ’ e t a - a - na - _ i i a - u i - a - a s-a- a - t i s - a -a- - a - i n-a-- a ’ a - t a - a - a - ma- a- La- - a - s - -a a- _ a- l t ’ ‘ a- a- ’a- ’ S c- ’n ’a f i ar i 1i - a - i ca-m ii -the an-ta a- I - a -- Tm it -a a - n-i a- 01’ a- a - ma - ’t a - a -c a- a -f n a - i- a--a- An i n t c r n a - l not’a-a i r stein term i nal axed - i l l r r o t a - a - c n l s a - lpn a -r - p r i m I u - nsa nia -’a - t a - ’ m r a - ia - m a - I lI t ca- a- a -a- sh m m a - u- a - a - x a - a n i l n aa- - it ma - a- ’’a - a n ’ m ’ - ’ n a - t - a- ’a-’- j - -‘ a i -a- a-a - t a- ‘a a-t u ia n ‘aa-’ a - a - am a- - - a - a- a- a- t la-’ -it t a - -re a - a - a - n mc ‘a- - a- a- a t ta- l’ t 77 or a- ’ a- I It b-a certa-iralv IV it a - na - a - m a m 1_a - a- a- ’ p a r t UCL a - i a - ni - a - i tic - a - Ca- m a’ ia - - a - m a - i n t a e r f m a --u- s e x i s t Lr a a - j n a-a - I n a-u m - ’ - a - t - t a- 2 5 h a - a - a - - a - a a - l o r a -- r i t a - a - - a - i En u ’’ i f l t a- ’ r c - a - r ’a - n a a - a a-- ’ L n a - a - i of a-” a - a - n a -a - - a- ia or -icr an a - i e ’ ’a- a- ’a - -a - m a - ee a -’t c - c l ’ a - a - a- a - i a - a - l ma- a - - I- a-i a - a - i c s i w s f r ca - m a - a -Jut r m a-n a-’ V a - u - a-l 1 76C a-a-- ’a- a- t a-i t a- -a-T i ni-a- - - ia- ‘a - a - t n ia-a n a-’- a - - ” ’i u e r a- J t ’- ’ m ” a - - t o a-I’ f l a - a - i a - ia- a- fl a h a v e a lea- t i l e - 1 2 a- a - - rk s u c t s - ’ a- t i a-s Ia- ‘a-’s a -n a - a- a- a - - a- - a- a- a- - a - a - a - a- a- m a - C ’CCSS a ra-’ t ia - ” m C c C S S Lb i e a- a - S a - a - is a- a-a- - a- c ’a i i” a- - u r e rn ’ ta-a- a- - ta -a-’ a - a - a- ‘ a c l c r l y ‘ a - a a - ’ i s a a a- - s a - t a - a-f a- -a- a-la-c’ ‘ a - n a-a- ru - t - a - is C I a - h a -’ 1 1 n-a- t a-- t ‘a - a - a - a - a - i ta - a -r W a - I l wall a - ’ t n u n a- s t a m - J n a- ’ a- ’ n t a-j a I’ ’ r i - a a- I t t a - n a - S a - a- a -n a- aa -a- a- ’ h a - a a - a - i s n n a t i a- c - I I t a - a - t - a-a- a ma - ’ C a-c’ a - n n a a I am aa-a - a i l a- r’ ooccls ’—a - a n a - a - I ta- a - a - ’ a - a - a a- - l i l a - v a - ‘a -h — a - a c e a - a - a - c - a - - a - - ’a a-a- a- a a- - a a - - i — t - the ‘i’ a u a - f l n a a a-- -c t l a-mfl e ’ ’ a - a a - a - i l a-i a I a- p a-r a - a - ta - i ra - a- e r a - a - - a-- a- a t a - a - -a-c i’m a-a -ta-a - a- a-a- 5 I a- a- a- a-f a-na-a--n ‘ ‘ a - I t- ’’ - lam a- a- ’a- ” ’ a- a - - ’ i a-- a -i a - h a - a - - a- a - 1- l t a - a - rra - n a - a - - i- a - s a l a - ra - q a a- a p a a-a-a- ‘a-a- a-a - i l l a ha-- a- ia - -a - ’ ’ - _ ‘5 a- a- a - a- — u p e r d ia - a- a- ç a- - -a-st a-a-’ a - ’ a- - a-m e cua -a - a- t r a - a -- m ’ a-a- - tia- c - use of t a - a- -- - all- ’ a - la - a - a - a-n ‘a-n a-- n n u t a - ia - - a - na - -a a- ti’ ‘ ‘ - necesn-a -a -t a - - - n ma I r a - l O T a- ’ r m l taa-c - Ira- ra - t a - iy n a- a-S i a-c An constructed a - n t n a - a - g a - n - a - - a - ‘ -a - a - t o a - It em otive a-u ’ ip r ce am a - ’h is t a- solcl’a-’ wa -th c a - xast a - na - q m r a - t a- t a-o s a - a - a - t i n s u c h a-u w a - a - t n a - t ’ may a - a - - a- a - a - 1 ia-ed a X 2 r 1 a s - a - a - t a - l a - ia a-- topi cs a- uott-a- 1 a - a - a - i a - l a - er a - a - c a - a t i a - r m on a m a - a - - a - o si mapp i ng ca - ill a-X a - i a i ’ l c s a-a- a- — a- - e s - a -- r a- -tU a- ’ i m l l v d i s c o n n e a - -t ia- ’t m a -n d when flow t i t a - ’ i r s i a - i n a - f n a’- a - aa - ca’ - -a-a- a- a- a -a- ‘ ‘ a- a-a- _ _ - - _- - _ _ - - - a- - --_ - - - - — 39 --‘- a- a- — PACKET RADIO NETWO RK PDP 11 ARPAN Ea -Ia - SATNET TIr’ - SIMP A A IXL U 1 PDP9 PDP9 UCL S PDPT1 r Su EPSS A SATNET S E ARPANL’T interfa ce I-’ig EPSS 4 2 in terface interface Pa-i r a -i a-i a- ’a - na - n a a- ’a - a - r l t a s- a -n of Neta-a-nrxs a-nd C Ta-tcways Ac a-cessa-bie f rai’n La-CL by M a- id— 1977 - a - - a - — a-_ -V - -a-a- -—--- —- - a - - — a- a - 5a- t- - i n C a- a - l a _- a- - a- ’ a- - a - a - - - a- c-ama - nr a - a-a-ac - a-’ - ci ma - a ‘ a - a - a - -a - i a-a-a- a - m a - a -I In a- a -I cm - a- n a - am - n’ v ’ r a - ’ a a- - a- a - ray a a ’ n p a - m a - ta - ’ n t a U m c a a - a- ’ a - ’ ‘ -- a-S I t c a- a a a - i f a - c a - a - a i um ’ i ’ aa a - a-h a- l ’l a a- a-a a a-a- ta-I a- a- ua - a- - a- i a -a-a-A ‘a - a -a-a- a - C a- i i r a- I t nag a- - - - is r c a- a- ’ I a- a- me I- it a- a- a-a-- a- - a ’ t - —- - a-- -- - - - - a - na-a - - - ’ a- ia - - - a I I - a- - a - - a - m - - a- ’ 2 5 a a- I ’ ‘a-l i t St a- C a- -—-- a- - a - c i a - a -- a- a- 2 ‘i a-- ma- — ta -a-a -’nca - at I a- - a- m a - ’- a- - a - a ’-a- a- - ’ -- - - a- a- a -a- ’ l’s a- a- a- a- a- a-a — i rc a - a - a - a - a -aa- - a - mt s Ir a- ia - a - a - a- a- at ‘ i a - a-a -i - a - a - a - a a- - S t a - - n a - s a - a - l t ’ i i in - b e i a-a-- — — a- o - a- a-nt ‘ ‘ a- a- mc n a- ’ a- a- 1’ a- a - a - a -i ‘a -a -i ia - i a- a- ‘sa - a-J a- L ’ C a i- a- a ’ a - a - r a a-’ a - a - - a a - a - lner r t r a - t c a - a a - l s Ofl UI ‘ c a- I a- ’ ma - ’ g ’-a - aC a - a- a-ma- - - — -- - ann r - e a- a - v e u i m a - u -L a-i t i V l t v a - m ’a - ’ I ia - a - c— aa -L a- a - a - a - c a - m a - a -a- - ia - a - a - c of ‘ta-- a- ta a- a - ’ a i a - i l _ a -’- ’ a-Th -- ‘- - - - r ’ a- -a-t’ a- s pa- ’ ’ a-a-ni ma- m ’ - ’ ’ a-a- a - a - - a ’ - --an a -a - a - i -a-ca-c a- a- h a - a -n aa - O t - a - m h 1t ’ a-a- a- ma - a - a a- ia -a - - -a - a - i a-t a-p la - - a - - a- i t ‘L a - a - a - I a- ’ - - s a - a a - n i a - a a-- - h a -a - a - - ‘ i a -a -i ‘ a- - - a-o a-a- a- - a- a- ’ it a -a ss a-Ia a- a - m ’ ’ n l u a - ’ a - s a - t a ‘ - a - -a -a- a a - a-a -a-a n’ a- --’ t a - S - t a- a - ’ --a CE a - a - - a - a - Ia- a- t a - ia - a - a - I a - a- a- a’- - ia - e n-aa-a- a- -‘- ‘ -a- ’ - a- - na - a - - a - a a- ’ a- i a - a - ’ a - i a- ta- a - a - a - a - I - ra m a - a - a- ‘ 5 ’ a - a - a - a -a-c - a-a a- - a-a- — - - a - - a ‘ -- ‘ a - a- - a-c a- to a - a - t t a - a ’ a -a a - t ’ a- l i u i a- ’a - 1 T ’ ‘ a - a - a - a - a-i s a - ’ a- a- T T -ia - -a - -er h-a- a-a-- a l a - ca- c’ s - a - a - t i n a - I will at a- is a- 5 s- c t Ca-i a - m t a- a-a -c - - C aa ta - a - ’a -2a -a R 5 a-i a - h T RPa -Na -ma-a-I a--a - S I O N C O N T R O L a- P R C Pa-a-’C-I Ia-i t ru d u c t i o ra - 5 1 In Kirstein 1976A we described our work on one of t h e p o s s i b l e procedures for e n d — t o — e n d conc a ta- a- r ated networks is put into each host cornrn u a-a- c a t i o n t h r o ug h in a-ba -is procedure special software a-a-a-a-u a s p e c i al I n t e r n e t P a ck e t IF a-s defined to allow a-i u n i v e r s a l h o s t - h o s t p r o t o c o l to be a d o p t e d across t h e d i f f e r c r a - t n e t w o r k s function is a-ca-ja- ired in ta-a-is a-a-h e gateway At app roach ca-owever p r o t o c o l mus t be comprehensive a minimum ta- a-c host-host Cerf 1974 Cerf ari a-a- K a h n defined a possible protocol which has been much extended Cerf 1976 Its embodiment is here called t h e Transmission Contro l Program Ta-ia TCP TCP a - a - as ueofl ia-amma ’1emented by three groups at BBN Stanford U C- ‘rf a- a-a-a -T a-a- a - C a - a - A f i n l r e p o r t of T i a - m a - l i n s o r a - t h e findings of th e c xa- c r im e n t s ia-a- in p r e p a r a t i o n C e r f 1 9 7 7 The UCL i n t e r e s t W a-a-a -are i n t e r e s t e d in e v al u at i r a - i n the TCP has a - a -a -v a - a - r c a -ia- ori g in s the suitability of thia-’ protocol for d i f f e r e n t i n t e r net w o r k o p e r a a -i a environments we w i s h to be able to propose simplifications and Gateways of Fi g 2 2 use the I n t e r n e t P a c k e t h e a d e r 7 t h e r e f o r e we ha -a-a ve a p r a c t i c a l need to nave an efficient ia -a-ple a-a-ac a- ’it a t i o n of ta-a-ic TCP a-c pass improvements traffic Fi a-ally the PDPI1 S AT N E T ove r SATNE T In a - ection 5 2 we w i l l discu ss t a - a c experimen tal tools These for Tia -a-a-e Stamp ing packets at critical points a b i l i ty to cia- a-a-na- ic ’ para-uii a-aa -ters remotely ana- a c a r e f u l des qn of the e- pe rimental controller and traffic generator include methods Ira sa-i te ways resolvea-i a- the experiments raised na-ore q u e s t i o n s t h a n For th i s rea son we also resorted and simulation methods ta- a- a- tacey other experiments In one different sot of experiments we a-d the b e h a v i o u r of a-a-ra- individual packet pass inc between toe e x a - a- ra--a -a a - a - - “a-P s ‘ a-ia- uc ua v i our Na- two rra a- a- a- ta-a-is way we w e r e a-au m C to from delays encountered d i sc -r a -t a - a - r a - g i c end—to—cru d inside the Data Transmission 4a- — ‘ “- E 1 L L ’ll iL - ‘a- a-mr ‘ 5 2 La -a - a-a- a - i a- a-i a-a a- a - c a-a -a- a - a-a- a- ia- -- a-a-a- u- h a- n ca- a- a - r a-ia -s a -ia -ia l a - a-J a- ’ - - ‘a - p a - -ni a - a - i ” a- a - r s 01 ta- a- S - - ’ pa- a- a-a-a- S a- a- ‘ ma a- n a - -- a- - i n a - - r la-c j f l a-t h were ‘ a-g ‘ a- a- t a - a - a - a- a - a - t a a -a -ir a --i - - -a - a - a -’a - - a--a- a-a- a-a- a- a - L a - - ta - a -rca- O -a- t - L - a-n a- a- i a - ta’ a-a-ara- t a t ’ i a - pJa - e i€- a a - t a tj a - a- a-’ a - _ a a-a- a -a- a- a- aa- t f l n n - a-t a- - aa- ra -swe r a--a-ia-i Ii i t a- ’ ‘ h a - ” s ’’ a - ‘a- -a - CS Ia - a or - a-a -e r a-a- 1a -c- -a - ram ’ a- - ‘a- a - a I a- a- H- a ’ a a - i ay c sisc a-ul d bca- a-t a ‘ a- ’ a- ou t a ’ a- L a a-a- a-a-a- a - n L’ia- la- a - a - a--’ H ’ N a -J r a - r i n a ’ ’a - a - l a- a- -‘a - a - t a - ia a-’- a- c u u sa - a -n a - I i ’i i ’n t aa- a - l a - L a - t a - S r a- c a - a - a I a -j f r om tIa -a- ’ T CP -a-a-i a s ” a- u sa - a- ’ a -i La-a - a - a - da - a a - a - ’ aS e a r l y in Vci ’ ‘ - ‘ r t a a - t r a i r i t S - 5 a - a a - Ta - La- t a- ca- ta-a-’ e ff e c t s ve a-aoss 1 a - a-”ci a- rs a- a - ‘a -r a- a- t a -h e -a s s Sa- sa-fl a -a -a - st na-y the i a - - a - a - ‘a-a’ic a-c rs cn r ied were to be t h e i -I P n r a - ta -a -a o l The a - t a -a - S I C fio a- c c a - n a - r d I ca-a- ’f y f t a - a - t a - a - s t a - f na - t a - s -am a- a - a - u t a-a- n a- ’ a-m a- a-a- -i ’ i i i i c r a - L a - a-a -’a a-nc- -a-a-ret i c a a-a- t a - t aVe ‘ ‘sa- ’’ t a - a - a - a - S t ’ i- a-a-a-a- e a-S a- ia -ca - -‘ f a’ Lfl dow a - a z - is a - i ’’ ft r a c - a - a-s b e in a -’ i t h e m P r a- ma - - a- ta r ‘a -a - i of -mma -’ a- ’ a a ’i a Fa-a- a - a - a i iv - a - i a - a - ’ ’ -a- ’ - ’ a - a - a-a- a- a- a a - a - I f ’ma i-a- i aa - c a - n a - r H I- taa - a - a a - a - -mr a- a-c a - ’ a - a -j t ca- i mu a - ia - a - a a - c ’ r a - - r a -j n a - a’ -- C 5 a- ” t S a- a- a - tm c ‘ia- a - a- - a - a -a - aI -n ’ ’ ta- a r a- b a- ’ ala- a- a- a - ’a- ’ a - i a - a -’ a ’a -ma - a - a r- i r t a- c a - a L a-c t i - ’ 1a- r- t o - - a - a - i - - ‘t ’ a- - a- ” - ‘ La- la - Ie ca- a- a - s u n - a-I a - a - a - a ’- a - a - a ’-r a - a - a - ea - a-a--”a- a- a-mt a - ai -“- _ a c- -r- ‘a - r r aa-fla-a- - p a r a- ‘a - l Uaa - it a- a- a-a-aa-a- — ‘a - a a-C a- c n i v e a- a-a- ’ th a-a- data - ‘ a - -m a - a - s L - a - a -a - a- ’Ia- ‘ta- s L l ’a- a- - ii H a- a - a - - a - a - a - a - a -a- ’ a- a - f a-a- a- -a - - - E a- a - i ’a -a - P l C f l flta -’a- a- F - a- a-j nn’ La - ’ ‘ t h a- a-a- i I a - -a - a - a -ri ca - a - s a-a - r a -a a- - ’ a- a - e t ’ a- ’ smc ma a - is-a-a- ” - aa-a-a-’a-SWe r’ u a - St u i a’ ta - a - a- f - a- a - a-a- a-a-- a - a - ’ a - I c ’a-T a- a- ’l - 1 a-’- - a- a- l a-a-a- ’ a-aa- a-a-’a a - ’ a- a- i a a - C a- r a- - a- a a - a ’j a a - o d a-a- ’a-a- ‘a- a-i- Ca-a- a - a - -’ t - j l aa - a -’ - a-’ a -a -a c ‘a- ’ ’a- a- ‘a- a - a - ti I a - h a - a - a - a- ma - a - -a- a- h a - ‘ a- a- C a m s ‘- ‘ a - a - a’ a- c a-a--k a- ta- a-a - i s o a -ta - i a - a - - a ’- a ’a - I a - a - a - a - a - a - a- a- a-a- d _ a i a ’2_ a-’ ’ ’ a-a-- ‘ - t a- ar e ‘aa-r a- ha ’a-a- a - a - -a a- a-a- t a o - a- a pa -m a- a- a - ‘a-u a- a-a -r y t a - a - C a-a-’- a - - a - c a - i h o r s ’ - of J f l a - aj a ’ a a - - - sv 3a - -” a - t a-a-i c a- - -c a - ‘ a- ’ a-a- a- a - ’ a - ta-a - s ’ a - l n- - ’ c t a - - a - a- c a - a-h c a - a-a- c-n a - -a - - a- ’u a-aa-’ a- a -a -l -a- i a - a - a - ’ s a a - s a a--a- a- - - - a-a-a-ca- - a-a- ’ a - t ’-Q s -‘a - I - ‘a - - L- a- a-i’ - ’ a- a- s i a- a-La-’ p r a - a ’ a - - l n ai- a - a - a ‘a-a-a- c l a - na-a - l a - a - S r a - ” t€ js i a- a- a- at ‘ a - I a ’ - a a ’ t i - a - hi a- a- a-2 a- a’s a- c ’j r ’ j a-sa -a i ’ a - l u a - a - i a -U a - c ’ a - -a-1’ t a- - s F’ a -t a- i a - a - a - a-a - em ’ ca - a - a n i fl a- I I a ta- -ca a-- - a- - I s - a - V a- ‘a-h a -a - a- a - - a t - a-a- a-a- a -a- ta a - a a-- - P C - i P a - v ia- a- r a - a a - t a - a a - ‘ am - a - p -- a- - -‘s- a - t a - a- - a - a - n J “ L sc a - a-a- i’ - ST -a- a a - - ’a- - in - a - t ’n aaa -u” - u ’ a - t a- a a - ’ a - t i S l S L a l I t ‘‘ ‘ a is a - a - r c a - a -ra- ’a-a i a - a - a - a m a - i a - a a-- a - a - a Ca - ’ a- a - ’ fa - iatS a -a - c a - I ma ’y ta -a have a-a-’ a-’J ’C ta t - ” r a a- a-a- a- a-a- ‘ ‘a- a- a- ’ a-a- a- - ----a- - _m_ — ii -- ‘ ‘ a - aa - S a- a-ze a’- a -I a -a - i Lafl p a- a - a -a- ha a- - ‘a- more uy cia -a-a- pro pa-a-a - r t a-es of para -a-a-a-aa-ta-a- r a- a- it i a - i a- ‘ a- a- L a-P If a - na - a - c slow r a t e t a -a -t o n a-a-a-La-V le ia -a - a-- a - ms alt u a-a-m th c ar — t a - c a - i a- -I t p r a ct i c e i n a- Xi a- ’ t a- a- u t tO i r a- - ia - c a- a - a-c F a-a-a- a-a-a - a - m t paramete r t a e a-i’cp a - a - a - Il a- na - a - a- ” n ’a a - C ’r l m e nt s f ra - i g a -t e a - it a t i a-maa - i a - a -a of a - a - ta - a of a-a- a - a - t a - a - th tja-a-a-e p a - a - cKca - a- set t o j u s t a-a-ave ta- - a a -a - a - t y of a-a-t a-a-a- the p a-is -k e t s a- Iii ca-a-ui i a - a - a - a - a - m a - ”a - c i i f i n i ca-c ’ re t ra - a - a - a - a - a - a - a - a - a - a -j on ia-i c 8 a a-Ta-J ciim ’Ia-a-sT a-ire a- ra’a-pi ementat iom -a a-t a-a- ’ a - J e a - ’a a- t or a -a -a -t a i a - a - c l u d c a - i i tems s s a - a - ’ a - as cm ’a - o i c a -a-a- - a -f b u f f e r i n g Ba- s-’a - lt — a- - a - a - s i i a - - a- a - i a - t a - j f j a - a - I ’- car v ’a r ’ ia - ’ a a - s I a - abut w h i c h L a - r ’ a - a ia ‘ a- a - l a -a- i a -n a - a - a - C O f l S i a - l u i a -L 1 t 1 5 a -n ’ai ’a-a- a-a-a- i a - a -a- s i n e a-mra-a-i - a - a - a’a - t As of C a - t ’ia -a- La - a -C ha - i n g ia - a - a - a - itat ia - a -ta - ‘ ia - ia - c ud a -a - P a--a - n a - e r ’ u n d er st a -mr ad a- a- aa-a- ‘ a-a- a- a - a - a -’ a-a-ia-’a- ’ ia- ci a - - a n y ir forma-ta-or a - a - u t s t a - a - a - a -a d- na - i t a-u study Cia -a-a effa--cts dat a - a - a- i a- site set a-a-a a- ia- a- sa - a - t a - a - a- a- a - a - a- ‘a- a - e r a - p a - r a -ia a -aLta -rS a -L i la - a - a -a-a-a- a ta-he Isa a f r a g aa - cn t e f a-he a-a- void t wa - as L a - a - a - a - - a- wa - a - a-a- ’ Wa-I S in a l r e s ul t a - a- l’j a-t iUa-a-a-i ta - ’a- riai 2 a - a - t a t La- a- a- uy t e s a-a-a- t U is a - a - ’ a - m ’a - a --a - C a - a -a -’a - - v a - i a -a - a - a - e d a- a-a-a- ga -a - a - r a - a - ta -a-”’ a - a- a- may have t a-a -a - a -a - sir a b i c a- i a- a- a -1O a - l a - r m a - m a - - a-’ t S i Ze ‘Ia-is a -a -n a-- a-a- i dependent egy wa- i a-h 90 a - C a ’a - c a - C r a - a - ar - ma - t a - size ta - i r a-a-ash a- i a-a-a- ’ ‘ a - a - s oc to pa - ’a a - a - a - e t s t rip ta - a -- a - a - ’ I a-a-ia-’ L a e a - a - a - a - a - -aa - ‘ a -v it a a- a- a- i-a -a- a - a - a- ’ - a s e s a- -t i ’ a-t a- a- a - s a - - s a - a ’-a - a - a - 1a - a - a - tsa - cs’a-- u t n a-ir a -a -cr i a - -a a - c - i r a - - 3 s e a t r e c t a - a -a - a - a- ’ a ta-ca-a- A a-l aI- a- ta-a - a - a - i iSSa--a-n U a - I U a- ’ r t S I r a - a s the a -ne a- a - a - a -xi r na-a-s - -Wa -i a -a SI a-Va - ia - ta-a-a- a-a- - a - i a - f l tt’ a - a - Ja-T a - L a - a - a-’ va t ic pa- sa-es a - a - a -’ f a - r c a- At w a - a - c i a - a -CLlQ W a - Ja- L a - a - a - a- ala-a- a- t a - ’f l t a i Sta- a- a - a - I defines pact-act s Ia-”- a -t a - a a -a - a- a- g e n c r a-t c f-a r t c’ iv s i z e of -a-b s ’-a- ’ a- s a a-a - Ia - c a - a - v a - a - i a - ’ a ve ry ba-a- aa - i a - a - a - aa - -ta-a - a - a-a-a cons a- s- ’ i a- i a - a - i1 TCa-- ’ a - a - ’ t a - i a - a --a- traffic ta- h a-a- a - 1 l t a -a- t a - n r a - a - aa-- m ay of aC a - l a-a- C t AR P A N I - i l ’ Ot f i x es a-a--i Vi a-a- a- - UCL in ma- t r a f f i c and a-ra e th c of tr a-a-ns a -ia- a -a -a-iia --a-a- c h m r a c a - a - a - r a - s t aa-a-s b -am ec s- i u na- a - a - a - aa-a-a- a - Iea - a-a-a- ’ a ’a-ta -nea- ‘a ta-r- a-a-a-s I c a -r rca-t a - a - I aa-a-- ta-na- se timesa ’ a - a - a - fi e la - a - a - e a- - a- p- ‘r ’fLa - ma ’ a - aa - a - a - a - a - cc of ir a - ‘a-na -- qjven iS tn e p a - a - a- -a - a -a -t ra-a-- iated to important amoun t The of a -a - a -j-osc ca a-a-a- v ta- a-c a - a - ca - Cd f o r a a- ’ u l i a-a-a -a-s at one end an d the maximum _ _ _ _ _ 44 — ARPA iti 90 si ze p a - a - a - k a- -t ’ -- ‘ a- — _ - “ - — sa -a-a- ta - ia - a - aoyt a -’ Choice of acknowledgement a-h a- e r at I a-C L It is pos sibi-’a- aa-i t a r m t e g y either to force out separate aicknow leciqem ents ACKs for each TCP packet a- or to p i ggyback ACKs on exist in atraffic in t h e ca-ta-a-sr direction where p ossia-le Tic choice ol’ strategy may enh mn ce d e p e n d i nq “ a -a - t h e is needed t ra -a-’at u r e o a - a - h a - a-a- La - n a-c of a - a - a - ia - a - na - t a - el t’ h -a- t r a m f f i c Ia - i a- La- e r a - a - a - e ’ a - a - a - co a -a -a - a d er ab l e aa-a-ay a-a-a - d caa - a - r ’ idc fj a -’ a- ’a- t r a - a - a - a - s n i i s sa- L’ - m a - s - i w i a - - t k i LJC L chose a stratca-ly a- per fa-a -i ra -a -a -a -a- cc ia - a - an ACK but a - a - ta -s a- COa -a-a-a- ’StCu 1 a- ’ ia - a -” i - a - a - H a - on e cta -ccica -a- t h a t t h e r e w a s no p a c k e t ised i n f o r r i a t -a -n p a-a-nd ina-ca- a n y w her e ba- ’ f o r e f o r c i n g o u t a-ia- a- - ALa-a- may still a-ri t a -a -p a -a-Ia- a-a- t h e tw o str cit ca - ia-a-a- l u’ a--a-a-”a-re t ’ OflC s e t ti a - n cr up by appropriate traffic situations TCP b u f f e r i n g strate -a - a-y iii As rica-ted a-ia - a -a --a- a-u t h e p rota -a -a - ol p laces great importance on ti - ta - a - r e l - a - ta ica - a-a- s-a - i a - ip b e t w e e n window a - i a - a -- a - a - ri d h u f f o r a-a- ’- m a - st r a -’a-a - i n a -s m e a s ut ’ a- - a-a-s a-a -a-’a- t s a-a-s I a-c opted for a-i fixa-a--d w i n d o w a - i a - c t h e y il a--a -o a- -a-Led for a f i x e d b u f f e r str ate a -j -a - The b u f f e r s t r ’a t e q v a - a - h a - a - e ra - w h i c h p r o v e d f i a - a - - m i b l a a- - en c a- u sa-h f-a- r na -st ra-’ a-l ’a -I lr e men t s was to m a i n t a i n a larqc ’ a- f r t ”-r - ‘a- a- a-’ni a-h fixed sized blocks could be chos€ -n a-is ‘ a-a- a- i a - a - o f f e r I a- Ca l neoa-ia-a-’cl Ia-aC- - a c e for w- a-s crabbed according t o demand f o r t r a n s m i s s i o n ta -a s - C e x p e r a -r r a - e n t a-v a-al a- abandoned this tende d to occur in the UCL a - and reception was cIa - oa -a -a -ra- so as to maintain at least one outst aa-’a’h ra- g f se rious space l i m i t a t i on p r a -a-I’i h a- ms - t a - a - a- - a - a - r e d rec eive n u i a-fer - em -m tic ’riments with larger window si-a-a-a--s a - is t a - c r c wa-is no ncceFsary a c a - a - n n a - c i a i a -n of n - a - c e a-a - -a’ ra- - - a-v a-i l a b l e - a-a-m t a- - n a - i a-ia-a- I marm y a - n r i - n m n a - t a - dy of a- A l -P A a -f i y a- I a-cl the a - h - a - a a-- t a-on I- ’ Lna ii y t i a- b e t w e e n th a - - P r O l l c r t s Lea - I pr a-a-t a-V ocol a- ca -u s e d by cho 13BN and ’ a- a-’a -t i a-na-I t h e ama -mount of buffe t ing strategy i a-a-” led to a-n w h i c h a--a- a - - a - ’ a- a- h ’ a a- n ’ a- ’ — a - - m-n a - a i m i n a t i o n of c o n s i d e r e d a-v t h e s i aa -alt I a - mt i o n 4 ta - orde r subncc t a- i n T h i t r a r y w i n d o w s i ze I a- observe the Ofl€’ a -a -ca -i i d a - a c t i o n b e t w e e n TCP and 1 ta - t a - a - v a r y t h e ARPFI pa ck a-a-ma- t a - p a - ’ This is a re ady -m ade contro l for examinin g the effect s of subr iet flow con t rol on TC P performance and Via -a - ’ ver sa Type ARP A C a -mCke tS are subject to ta -ca- f u l l ARr In era-cl—to—end flow control type 3 to none 3 ma a- a-a -ica -c ’tsrun t h e r i s k of t yp e i a - a -wa - a - vre a- being discarded at any poi nt era- -a-a-cute i f experienced C o n f i g u ra t i o n s ana-l e a -ma -od s 5 2 2 it T h r e e bas a- c t r a f f i c s lt ’a - a - a - t i a - a - i a - s we re tea-c wa -s a - c a - a - i a - a - t ’a- i a- TCP perforta-a-anca-a- i n f a - a- a - a - a - t i a - a - n ‘I a - a -u s e La - a - a- a-- Ia-i a a- a - ie n ta - fa -a - a-l t i i f fe rent a -r i t a - a -a-a- a--a-a- i a-’- oo na- w a-a-a - r u s i n k and the e c h o-l o o p self i a - -a- - p were received on the sara -a-c TCP a-art not r e p l y beyond a-a-cknow lea -ia - ing i t echoed re ceived t ra f f i c Th e s e l f — loop optimum TCP m i n im a - i l a-i i r a - a -a - - -to g ive ta - a -c - source— a - ’- -a-ra- smn a - i t t e d packets a- a-r t wh i c h did F a- the echo-loop ’aa - a - l y a-n i s a - a - i a - c a t i o n of t h e V a - i a - i a - a - l a -S i t a - a - a - t t i O a - a - S i r a - of a ma - a - a - a -a - mt aspects of ta -a- the s e n - i a - a - ‘w’ a-a-a-a- u a r ’ a-a p a - a- a- rft a -ra -a-n W a- l i c a - In a-he source—sink case t r a f f i c was receiveca- on a d i f f e r- a - a - n t doi ng blockages are ‘a - v a - t a - a - rec aivira-g on ta -m e a- a-na -c ‘la- P a - a-s in it ’ ta - a - a - it TCP i s i s a - r a - ama - a - a a - a - i t tin ga-a-a-i d a-ca -a-na -a --a- a - a - a - a - a - a - a - a - a’a-- a r- aa-ts - a- a- I The a r e au t o n a - a t i c a piggy backed a - a - ia - a - a - no e Xt e n si v e tTa- a- it SipU la - ita- a - a -- ma-a-S ‘a-re r e q u i r e d This 1--s- op is n o t p o s s i b le howeve r i f a - na - c TCP is t a l k i n g to ‘ ia-h e a - a -- a - u r a - a - a - - s i n k a TCP in a remote m a c h i r e p r o v i d cc more nora -’m’a--a -ia- s t t a -ia - a - a -io ns i ’a- a - a - T i c c a - a - i n g f u l l and h a l f dupa-ex connectiora-s c o-a-a- a-a- C ha-’ it el f howeve r ena -a- ta-a- ea-a- a- a-ie lays to a remote Ta-ia-a- a a -na - r c c ’—s not give i n f o r m a t i o n TCP a-ia-a-a-a- and t h e echo- loop ota - a - t a - ca -a - s e t a -h i s c o u l d o n l y be ca - ta - a - ir a - e d a-a-y u s i nq an a-a-ch o c o n rie c t i - a - n in order to obtain data -a- us a-icle at a-a-a- echo connections were almost always used for experim ents to remote sites Benaviou was studied in severa--ti c ii ffcrent Fia -ys ica l configurations i l lu s a -r a t a a - a - d in i’a-ia-J 5 1 Two l o c a l loops we re used In ta-a -a-a- “ i n t e r n a l i a-a-a-p a- a transmission q u e u e a-m a-a- TCP whe n tha-a-y a- a - l a - a - c a - a - Swere ni ace d on a-a -e ach ea -i the rread of it were immediatel y transferred t o t n e r e c e i v e qu e u e I n t h e •a-a- t o a-a-a-c bona-lona- IKP p ioop ” p a c k e t s were a - ra - a - a - a -msaitteaacross a la -ca-a-i host interface A s i m ’ a - i a- a r- “ VDIi— IMP loo a-a- was t e s t e d b u t was a b a n d o n e d d u e to h a r dw a r e prob l ems a-nd an unsolve d lock—up cond L ha-n in ta-he Va -il RTP aT r a f f i c was also sent t o a remo t e si te us u a l l y Sta n f o r d although some ‘ a- a - a-a- a - a - a - - —46— src dst src a- elfconnection - loops IMP _ ‘ I ‘‘ lOOK b s dst se jfconnect ion a- Goonh illy SIMP 9 6K - a-LONDON TIP a- - dst - - a- 9 6K bps NORSAR ‘ “ a- ‘ src echo TIP - dst boo IMP loops ‘ ETAM ‘ - SIMP -‘ bps 100K bps P a- source-link b - TCP src - IMP 100K bp s TCP src dst echo loop Internal ‘a- I - src source—link a IMP dst ps a-_ 9 61 bps ’ SDAC IMP - ‘- - - 50K bps ‘11 hops ‘ — a US ARPANET - TCP src i’i g 5 1 PDP9 SUMEX IMP 1 50K bps TCP PDP11 d est The Experimental TCP Confiqurations - - - connections were also made wita hf-N In or der to examine the effects of frona- the main ARPANET U C La - two configurations rc sm aa -a--a - a - i a - - c i a - t a - a -i were s t u a r c u t r a f f i c was s e n t across the s tan d a r d N O i a -ISAI-t r o u t e nominal ca-a-a-pacit y is bps w i t a - i an a - a - ti nra - a-aa- -ne in a - a - ha - a - a - a - nIe a-a- a - va -’ d a t a ca a-- a-a - cia - v of between 5 1 -“iou 7 6K a-s-ps for ARPA h o s t — t a - — a - c a - a - ma - ia -a- - e ts a-i a- p a - n d i n q a-3 6 K ba-i t c f fc a - t on w h e t h e r s e i s m i c d a t a Is ha - a - ia - a - a - t ra - a nsmit ta - a - c i f r o m N a - a - RSA a - - This I ra-a- channel is clearly a-a- bottle—nec K and Or ’ne ca - ia - a - a - a - x a-’” a - a - ct d e l a - a - ’sma- t o a -u i up i n N o r w a y a n d at SDAC on t h e r c a -u r n Jo a -a -rra -c’ v Y- a a - a-a - ma - er con f i guration broadcas t was t o use t h e 50 ’m bps c h a n ne l r a - o v i a - i e C ta -a - r t h e satellite experiments capacity of 25K bps u s i r a - a -j PT a- MA to f i v e a-a- ca-a - c—wa y In this case the p o t e n t i a l b o t t l e n e c k is t h e 9 6K bps London-Goonh illy link To obtain figure s on throug hput a-a-nd delay one h a-ms to know the times at which various events occu r the c r i t i c a l being the point at which -a packet enters a-mo d Thus the basic tool interfaces la -a -a -va -am a- the ‘ha - P for measur in a - TCP beh iv ica -ur ‘ c ts a-a- Ta-r ’ esta-a-m’a-ae packet which picked up timestamp values a-it t h e s e - p o i n t s -‘-a -s packets might be sent to e i t h e r the f o l l o w i n g events h a -i d 1 2 3 4 5 6 7 8 a- sink process or an ca-a--ho ‘a -r a - c- a - s s -“ to be a - a - a - i owe d ta-a- Generat a- o -a- a-a-t source a-recess Transmission from source TCP R e c e p t i o n a t a-a-’a- e s t i n a - a - c l o n TCP Reception a-a-t d e s t i n a - i t i o n s i r a - k ’ ch cca-c ca -s -a- Departure ta-c a-’ b a - a - t a - n a t i o n ca- ca -o cr ‘l’ranssa-i ssion fra-ina I-a-ta- cc’ a - a a- iota - a -ma - s t a- ra-at ic-n ‘ia -P -lt source rCp a Reception at s-a-a - a - a a-- ce -‘ a- b o g gi ng - - - ‘ process a-’ - - ” -- - ‘ ‘- - - - a m offset - field to point to the first unstamped loca tion in the f eld the timestamping procedure was made By including s imple to code and operate i’ ig 5 2 The forma t is illustrated in -a- S - TCP h e a d e r Offset Stamp 1 Stamp 2 Stamp 3 ‘a-”Times tamp area Stamp 4 next time stamp ag oes h e r e - Stamp 5 - Stamp 6 Stamp 7 Stamp 8 Data Fi g 5 2 Timestamp Packet Format Throughput can be calculated from the differen ces between c - a-r r e s p o n d i n g t i m e s t a m p s in successive p a c k e t s d e l a y s f r o m t h e dit ference VIa- a- S between timestamps in the same packet measured directly Throughput in terms of the number of letters ri tt e d per second by the source process where a-i trans“ le tter ” in TCP terms is simply a unit of data which has logica l significanc e a _ _ _ _ _ a- a - a - a - a- ’ r s t a - a - n a - a - W I m a - e ra-a-v a- a - a - a- a- ’ a-I t a - a - S a- a - a - - a - a -a-a- a-a-a- a- ’ ’ ‘a- a- a’ - a - ut a - a - t m a-a-a-Ca - ‘ ’ ’I ’ a- a- a- a- - ‘ - a - a - c a-a- ‘ n sa- a -- t a - t a - a - a - a- h a - a- ‘ a- a - a - a- a -a - a - m ”a-u --i-a - a - a- ‘ a- a- ’ n a -a - - a - a - a - n-a m _ a- a -’’ 5 a - n a - a -- f a - ’a- a- a-a -i’ ia - C i a - a- a- a- a- a- ‘a - a - a - a- a - a - a - n ‘ a- a- - a -i a - a - a a- ‘a - a - i ’ _ _ _ _ j a - a - ta - y a-a- j a- ’ a-a- a - a - a - ma -a - m a - - a - n s a - a - _t aa- - a- _a- a- 1 n - a - -a - L a - t a- a -ce a-a-a- a- ma - a - a- ‘ a - i S ia - a - a - a - a - id a -ri of a- - ia - a - - h -- a - a - a - a - s a - - -c a- s a - f a - i I ’ ’ a - L a - a - a - a- m- t b a- a - a-im a - ha - s cf ‘ a- n a-a- a-a- cc a- ’ i’a -i ’ a- a- I I - a-m a - a -a - a - a - a - a ’- S- -a - a- a- ’ a - a - ’cd - a - a - s a - ta - -na - a i- ’ it n a - a - a - ft 4 a - a- — ’ i a- - a - a ’- a - a - - n - - a - ’ Ia- a - a i- t v a - - a- t a - b a - a - a - a - ‘a a - n t a-a- a - m b a - i a- a- r a- a- a- - a- nv a- a - a - a - a - a -i l a - a- - a- -a ab a - a-a- -a-na -n ‘ - - a- i Si ’ na - a -- t w a - i ’ a - ’ - a- a - a- a- a ‘ a- _ ‘ - a- _ a- a- a- t a - a - aa- _ a- ma ta- _ a - a - a- a- a- la - a- ’ a- a - i ” a - a - a - a- a-a- _ _ t t a- - i la - i a- s - ‘ - a - a - i t ta-a-a-it _ - ta-a- _ a- a- a- a- a - I - - n ia - a - a-a- a- a- a- C Ca - a - i a -- a -’a-a- ‘ a- ia - S a - a- t a-a-’ a - ta-a-c _a- te r a- a - la - - a - a -a - - ta- ta -a- a-a’ a- La- p a - r a - a - a’t- a- a -i a- - I a- ia- c a - a - a - i a - a - aa-- a- ma - a - a - 5’I _ a - 1 ’_ a- - i a - a- a- a- a a- a - a-i C 1 a- a- a- a-a -a -m a- 1 be a- ’ a- a- ’ a - _ n a- a - ca- a - a - a - a - s a - a - n a a-a- - a - - c- a-a- t a - i a -’ a- X a - a -a ’- a- ‘ a- - n a - a - I a - a - a - m a- a -a-a- - a - - I I Wa - a -5 ba - a - mi a - a - a - a- I’m t a-a-r t ’ a - a - aa- - ’ t 5 a - a - a- I a- a- a-a - a - ’ a - - d a -a-’ r l a - a - - ’ a - ’a - t a- - a - ‘na- a- a- - a-a- t a- Aa - a- a- i a - a - ma - a - ‘ a-b I a- a - a - ’ ‘ a - a - _ a - a- a - t a - -’ a-a- - m a - i a - a - t a -- ma - a - -i ha - -i ta -I a - a- a - a- a - a -” Ia - Ia - ’a - a - t iCa - a - ta - a - r a- ’ - t a - a - a-a - a- ia - a - La a-- ta - a - ta -‘a- —‘ a - a -- a - ’ a- a -i a- 5 ta - a - a - a - a - C ’ a - a - a - ia- a - a - a - a- a- t a - a - - a ’ - Ic na - a - t a - a- It i t a - a - a -a-a- m ’ F t a - a- ’ma - S ’ a- ’ a - S a - ia - ia -’ s st a - ’ri L a - ’ ’ a - _ a - t a - la - a- a- ‘ha - a - a - a - a - a - a -a- t a - a - am a -a-a -a-a- a- ia - a - a - ’ ’ - a - a -a - - a-a- t a- ’a - ’a - a - a - i - a- a - a - t a ’ - ’ i - a- a - - - a- ’ a-a-a-a -i i a - a - a - a - t a -- 1 jI m n a - a - y a - a - a -’ __ I a - a - a - na -S a- a-a- t a-a- a- r a- a - a - a-a- a- a- aa- a - n a - n - 5 a- ’ a- a- a- 1a- a - il a - ’a -’ a - t a - ’ C i a - i a - la- ‘ ii ” 1 I i-- a - n a- - a- -l ma - a - a -a- a- ‘a - a-i a - ma - a- a- a- a - _ a - a- -i a - a - -t a- a- a - a - ‘ a - a - A a - a - a a-- a - a - st ‘ a - a- a- ia- a- a - S a- a-a -a-Ct a - a - a - a - ia - a - mi ’ a a- ’ a - a - - a- a- a- a - a - a - a -- t’ a - a- a- a- ci a- n ’m’ wa -a - a- V a - a - i’ 1 - ” a - a- i m a - a- a- a - a- I t a - a - a- a - a - a-a-a-- a- a - a-’ a-a - -a - ’ea - a- a- -_ a - - a - ’ a- ’ i n ’r ’ a- a -a -- a- La - i a -a- a- t r a a - c i r I a - n - a- t a - a -- a-s a- a -I -a- a- i ma e a a a - a - m ’ ‘ha - i S 1a- a - a - a- -a - u n a- a- a - c a- a - a- V a- I Ca - a - a - a-- a - c i a - t a- - ’ t r a - t a - 1 a-a- ra- a- a - - a- ’a - t a - a - f l a -- L a - a - a a-- t a - i a-i 1 - -- a- ía- ‘a - a- 4 - a -c a - ma - ia-a-a-a- - a- ’ a- - ct a- ”t ‘a - - a - i a - a - a - a - a - a - a ’ a - a -m a - a - ma - S - tS ’ m ’ a -a- a - a - r c a - n u p a - a - r t M - a - nu a - i a- a- i ma- a - a - a - K C -t a- s - a - i a - a - a-a- a - na - a - a - ia - n Ia-a-a-’ t - -a-a - m a - a - n h t a - a - a - ca - - a-n ba-a- a-a- a - t a - a- a - i a -’ a- ’ a- -sa - ’u r ’ - aa - j I ’ a- a- n La - a-a-a-i -a- a - - V a- - a - a - I a - a -a -I _ ma - a - a- - n - a - a - On i a - 5 a - a -- f l t i _ t fl -a-t sa - l ’ a - i i n a - t a - a - i t t ‘a- - r a - - ’ a-- La- a - i a - t a - a-i a n s u a - a - p a-a-’ c ta - a - a - a- a - a -t a - i a - a - n - c a - a - a- a - l a - - a - a - f r a - a - a - ’ a - - c - a - a - i ai a- - a - a - a -SC a - ’ n - a - C a- a - a - a - a - ta-a-a-’ ca- Y t r a - a - t a i a - ’ L’ ’ a - a - a - a - v a a-- a - a-- a- - - a- ‘‘ t o - ‘a-N -a - a - a- ’ a - n a- a - ata - n a - a - a - a- t r a - n a - n u a - s a - a -- n a- na - a -- a - - ’- n a - I r a a- - ‘a-i a - a - a - I i - a - a - ta - a - a- ‘ aa- a - a - iS a - a - e ta -mi t— J o I c a - V a - i S O a - i a - a a- - l a - i a - f l a- a - a - a - i a - a - a- a-a-po r t a - a - a - a - m a - a- ca-I a- ma - ’ l a-a- ’ a- a-L a-a-t a-a- a-a - a - a - a - sa - a- a - a - a -’ a- a - a - a - a - a n- a - u t a - a - a - a - ta - t ea - A m t a- -In I a-a a- a - - ’ a - t ’ a-t - a a- - a - i a -a- a-a-a-a-a- a - r i i m a -’ a - - a- a - a - I ’ a a-a- t a-i a - ma - ’ U n -a - n ‘d - a -a - ’ aa - va- a - a - a- I ta-a-a- ’ a-a- a- C 4- a - ma - a-’ a- a - a - a- ’ a -a- - m a - a- ca-a-a-’ t _ a-a- a- - i a - a - ta- I _ a- a - - a - - a - a -a - m a - a - a - m a - a - fa l - fc l a - a - Wa- i - r -r ta - r a -r a- a- i t a - - a- - a-a- a - i a - a- ’ ia -a -s it - - a ’- ia -a- ’ a- a - a - ia- ’a - a - na - a -’S a - a - n -a ta- c a- a- -‘- a- ’ - a -— ‘a- a- a - a _- ’ a - a - - a - - a ’- ‘ a- a- ’ _ ‘ ’ a- a- ia _ a - a-n c a- Ia- ‘ ‘ a - m ’ th c a- ‘a-- a - a - s a-’ _ - ‘ - a ’ a - a - a -ca- a- 1 — 50 — t r a f f i c a- a a--a - a - ’ m a’a-- a a--a - a - a-a-a-a-a- w o u l d ‘a - l a - i s m e ch a n i sm e n s u r e d t h a t occur at the optimum rate wh i ch co u ld t h u s b e measured directly iv Parame t er Change Th is was a sp e ci a l pr oce ss w h i c h could re q ues t or implement p arame te r ch an ges fo r bo t h the local and remote sockets of a connection a n d th u s in theory enabled a TCP experiment side only to be run by one In practice it was only ever implemented at UCL due to space restrictions at Stanford v Data Logging All incoming data was routed process which selected tiniestamped intervals and monitored parameter through this pac kets at regular change responses as well a s main t aining a wa t ch o ve r b a ck ground information such as the opening arid clos ing of a cor a -nectior a - All these processes could be d r i v e n either or automatically from commands stored i n a s i m p l e co nt rol la n g u a g e Post -processing ori g i n a l l y in a command file written remained rudimentary intended that the TCP data captured onto magnetic teething facilities from a command console It was should be lo gged tape for subsequent aa -n a -a - lysis but ow ing to numerous problems with both TCP and the tape hardware and software this was never implemented and proposals of run—time data reduction for a-i certain amount were not studied u n t i l late in the Both timestan -ip data and the TCP packet traces were interva l at which timestamp packets were traced Subsequent project loa -jqed on to the line printer w it h t h e us er sp e c i f y ing the r e duc ti on w a s carr i ed out b y hand These limi ta - ’a -tions were never a serious prob lem as the experi — program never reached the stagc a - of generating mental dat a too large to handle were made by samp ling every of 100 the sampling Thus a very each point 5 3 All throu ghput tenth incoming F’or a f u ller packet indicatin g t h e u n i f o r m i t y few calculations a c t u a l l y Experimental and d e l i y of a-i volume of na-ea-isurements from a group the sample had to be performed for Results t rea tmen t of a- t he experiments condu ct ed a nd t he - — -a - a-” — - — -a - ‘-a- - a-_ a-a-a-a- a- ’ ’ - — ‘---- —---— ‘ ----a- a- -a- — results obtained the reader is referred to the TCP report Cerf 1977 This section outline s the results obtained and discusses their significance for the TCP in the UCL environment 5 3 1 The UCL Implementation A large number of experiments were carried out ira- the two DCL configurations of Fig 5 1 - internal loop an IMP loop A s a result of the earl y measurements undertaken a n umbe r of serioa inefficienc ies were revealed which were corrected although the DCL TCP is still inefficient For instance under our PDP9 operating system the new occurrence of an event eg I O interrupts is indicated b y a known glo b al flag changing state The scheduler at some later stage a- detects the change and passes control to the appropriate service segment This scheduling by flag causes delays for example even though there are packets awaiting transmission the data will not go until the transmit state flag is inspected in the scheduling loop It was found that in a heavy traffic situation the hardware data adaptor was idling for relatively long periods Although a simp le rewriting of the TCP scheduler resulted in a dramatic improvement see Fig 5 3 the adaptor was still underutilised The round trip figure s of Fig 5 4 for the DCL IMP represents an effective capacity of 26 35K bps across the local host interface which has a nominal rating of lOOK bps This suggests that eve n in the comparative ly idle situation represented here the DCL TCP is compute-bound Results in later experiments support this conclusion - - —a- -- a- - ‘—‘-a- a-’- a- -’—--———a-a- -a--a-’—a- a - ’— — 5 2- - 1250 1000 750 Delays ms 500 I Il I a- Il It Il IL It It - I t IL It It 25 Packet 0 Num ber W i nd ow s i z e 9€ bytes Packet size 18 bytes Thrash size 18 bytes Retransmi ssion 8 seconds Con fi gur a tiO n t self loop to IM P Sol id line old schedu ler Broken line new scheduler F i g 5 3 Transmission Queue Delays in the DCL TCP L a - ’a- a-l a - a - t a - a - I’ ‘I I ’ a- a - a - a- a- f a - a- ’ a- v a- a - a - a - - a - ia ”- a- a- aa- a- a - a - R a - - i’ a- a- a - is a- a-” a- a- ‘a- - 1 I i I i a-’ t r ina -’ a - na - ina’- s a - a - a - í a- t a - a - a’ - a- I - a- i a ‘ ‘ a - i a - a a- c i ‘a- wa - a- a- ‘a-’a-” a - a- a- a- - ‘ a - a - a _- i h In to c t m a - a - a - a- no ‘ a - a - a - a’ - a - Ia - ta - ‘ i a - c a- i ’ ta -a-a- - n-a a - i a-a-a- 1 a - a-r a- a- - a-ma-a-a- ’ - ‘ - - ‘ - “‘ a i t a- - m a - ia - a -a - a - a - a - a - a - m T a - a - m a - c- a - a -i da-a- a- j a - n a- a- - ra- -a - a- a -a - ‘ a - t n a- ’ a - ia- ‘a -k ’ i a- f a - ’a - a - a- • ‘I- a- a - a - a - a - I ’ a- a- a- ’j ia-a- f -a -a -’ aa- -a - a - i a - a - a - a- Ia - a - ’ a- a- Oa - a- a -- a - a - a -t a- a- ’ ’ ’ a - t ’ a - SI a - a - a- a- ’ a- ’ a - n’ ta -a- aba - 1 a - - a - t a - n a - a - i a - Iua - a - i a - a - a - a - s i la - a - ii a- ‘ a- a-a-’- f a-a-’ ia - a-aa- a - a - ’’- ” ’a- a - S a -’ a- ’ ’ IS - a - a - m a - ’ a-i- - ‘a -’ ’ I a - f f e ’ a -’ a- ’ a - a- a - W a - a - a - a - a- ‘i ” a - a - C a -’ a - Slfl a - a -ta -ot i a - a-c a- a- a- - a-c a- ’ a - -ma - c a ‘- a- ’ - a - a - a -- a a-a-a- a- t a - a- a-a-’ ’ a - a - aa - a- - a- a - a - a- a-n ‘ a - a- — - ‘-a - a - ma -a-a’- a-a-’t a - a - a-a- ‘ a- a- a- ’ a - r n-e r a - ma - a - a - a --a- a- a - a - a-a- of ‘ a- a - a - - L i l a - a - - a - a - g a -- a - - u t a-c a- -a - ‘a- ‘na -a-a- i a- a- a -I i o a- a ‘ C a - a - a - m a - a - n t a - a - i a - a- a - t a - a - a - aa -a- - ma- a a- -a a- - a - a- a- a-a -a - a - ai- a - I t a- a-a-cr n a -a- a- ’ s r a - c a - a - I a - a - a - a-ba-a - n a a- - r a - a ’- a - a - a - a - t a- t a - r a - a - a - a - aa- - a - a - t I a-’ 1 a - n - a- a - a - a - ta- a - a - a - se a - i ’ a - V ’ a - ’ a - a - a - i ’ t a - a - - n a a -a- ’ i m a -a-j i_ a - a - a -t m - ‘ - - - a- a- a-U a - m t a - - ‘ a - a -s - a- - a- a-’ i a -a- a- a-a-’a- a- a-a- I ar - ’ a - ’ ia - ia - a - a - I - a- r a- - a - a - t a - a- a- ia- i a - i a-a - a - ta - a a-- a - a- a- ‘a- - a - ia a-a-’a ra- a-a-a- a -a-a - a - l a -a- a - a ‘a- a - S a- a-a- ’ a-a- ’ W a- I ‘ - -a - - — - a- a- - a - a - a - a -i a - a - a - a- - a - m a - ‘ ‘Ca a- a-a- a - “a - a -a- a-- a-I - ‘- - - - a- La - a- of a- a - a - c a - a - a - a - a a- - a - a - ai t ’ ’ a- a-a- i a - a- a-h n’ a- a - m ’ m w a-a-t a-a-a-’ Sa- a- a- ta- ‘a-’ a- - ” a - a-a- ’ c a - a - r r ” n t ly a-a-a- a- a- ’ a- ’ a- a-a- a- a- a- - a-a-a- a- a- ra - a-W r a - t t a -a- -fl f I a- c ‘ a - ’ — - a - ’ - — - a- ‘ a-a- a - a - - a- a- a- ’ C S i a - a - a - s i c na-a-- a- aI- - a - a - n -aa-a- - ‘ a- a- a- a-a‘ ‘ a - a - t I r—a- a - a - n a a--- a-a- ’ — n a - v a - a’- a - I t a - a - - a - a- a- a - a- C a - m a - c r a a-- a - ia - a - m is a-a- c’ a- -a- a - a - ’ s mia a - a - t a - a - ’ t t a - a - iT a- - a- a - a - ca -’- s s ia - a - a - a - a-a-V a-a’- a- ’ ‘n - _ a - - - a - i - a - Ca- a - a a - t ’ a- ’ f a - a- a - a - et a - 1 Ia - a - a -t a -a- a- a- ma - a - a - -ffica - a - a a- a - a - a -a- a- a - a - a - a - g a- a - a - a -- ’ I a - n a-a-a- a- ” a - a- a- a- a - a - a- a- in a - a -a - a - n a - C- a -a - a -n - i a - a - a ’- a- ’ a - ia- a- ’ t a - - a -t ’a- n a - ’ i -a- a - a - a -t a - a - a a- -a - a va- a- a-m a - m a - a - ’ a--’ t ta - a -’ a - a - m s a - ’ - ’a- ’ a--a- n r l a - I a a-- a - i t l a -a - a- a - i ‘a- a - a - a - L a - a‘I- a - a - a - a - aa- -a - I - a- - n a - a -a - a - a - a - a - a - a -’ i e n n a - a - - ma - la - -a - i iy m ‘a-a- - W i ns - aa - - - i a- Ir a a- ’ Ha - -t r a - t a - a - v - a - i a - a - U ‘ a- to a - i a-a -a -a- a- - ma-m a-a- ’ r a - i a - f l t s a-a-a- a - a - n t C i a - ‘a - ’ a- ’a - n ma - - - a- a- t a - a-a- a- a- -a ’ a - c a - ’ of ACK a- ’ a - a - a-a- a - a - aa- -n - a- a-a- a- a- Jnt a - ta - a - aa- -a - a - ‘a- a - a- a- a- C a - a - n a - a - C a - a - t i _ a - i a a-a-v - s a - c i a - n a - i s a - a - a a- - ta - a - a - ta - a - ma - a - a- a’- a - a - aa -- t a- a- ’ - ma- - - a - ze a - r Is m a ‘ a- L ’ ’ o a - a - L a- - ‘ t a-a-a- a-il ’ CLa ‘ a-a- a-a-a--r 0 1 m e a- t ta- a - - t a - a - a - a - a - a -‘ 2 r a - p - a - ’ a- -a - a - - i a a- a- i a- a- a - a -a a - a- a - la ’ - i a - a- a- a- r c a - a - c a - a - m n a - a - ’I a-a- ‘ La - a - ’ a- a -a - a- a- ‘a-’t c5 id ‘ a- a- - a- a - c a - i a- a- P a - - 4 - ta- a-I a- a- a -- a- a- n a- a- s t ’ a-s a- ’a - a - ”aa- -a - a - a - a-W 2 ‘ a - a - v t a-_ a - a - ta-i I - i a-t’ ‘ a- a- -a a-i - — a- ‘ a- ‘ - — a - a - m a - a - c a -’ m ta- t a - r r a - a -m i a - L a- ia- ‘ a - a - a - a - a - a - a - fl t 5-a- a- a-a- ’ a- a- a- a - a- aa - I a a-c - a - a-Ca- ’ ’- a- - ’ - -- a - ’ —a -— - -- a- a In ternal loop 16 14 ’ a- 12 0 1 2 3 No 4 5 6 7 of letters in window Letter size 18 bytes Retransmission 8 seconds Solid line echo loop IMP b 8 Thrashsize 18 bytes Broken line self connection loop 14 12 1 0 o g t t 8 ° 0 ’ a-a- 1 Fig 5 5 2 3 4 5 6 7 No of letters in window 8 Throughput for one and two connections - a- _ _ 1 - — - - — ‘a - — a - ta-a- - ‘ a- i ’ Ca - a - a- ’a- tie a-a-a- 5 a- a a- — ma - i a - a -a- a- ’ a-n a- a- a - a - a - a - a - aa- -t ra -u - “ p ic a-et a-a -ta-a- a- ia -a- a- -‘a- a - i ’ a - I a-a--c a- a - a - a - a - ’r - ‘H A l a - a - a - a- na - lu Iv a-- a-’ a-i a - a - la -- a - a - S a - C ’ a - a- a- - a- a - - - a- a-m i - I - - - a-a-a- a-- a - a - - a- a- a - a- a - a - t’ a-” a-a -ma - -— a- a- ma - a - ‘ a- a- a - a - C’ a- t a- a- a-’ a- - a- a- a - a - - va- - ‘ a - I • - Lii a- ‘ a- - - - - a- a-a- a-a-i a - a - ’ a--a- ’ a- a - a ’- I a- - a- - a- a- a-a- i l0a- ai_ a - a-C a-a- a-a-a- ‘a- a - a -a- a - a- a - a -a - a- a- a- a- a- -a- a-a- a- - - -I t ’a - a- ‘a- a - - I va- a - t i c I i ‘C a-’ - a- a- r a- 1 a - a- a - a - la - a - ’ aa - a - a - a - a - i a- - a- ‘a- a-a-’ a- a-m a -a- ‘a- ’ a- ua - a-a- a-i r a - a - a-S ‘W - Im l- ‘ a- a - a - a- a- Ia - a- a-a -a- a-a- a- ’ a - a - fl Ca-i 6 1 a- -a-- a-t a- ‘ a- T ’ t S a - a- ’a- a- ’ a- ‘ Ia- a-a-Ia - is a- a- a- Ca -i a- a-a-a- a-a- a- - a - C - a - a - a - a - ’ a- a-rn a- m a - a -a - ’ a - a ’- — a - a - a t -a- a - m a- a- ’ a- a a - ’ s a - a- a- ia a- a- a-a - a - a - s S i ’ a- ’ r’epa - a - ’ _ e a- a-a- ’ a- a - m t -a S a - a - - I a- a-a-a- a-a-’ a- i f a- a- a- a - a - ia a-- a - ’ l a -_ a - C a - m’ -c a - a - I C - a - a - a - a - a - a- a - a - ‘a- a- ’ a - a - a - a- ’ 1 1 a - a - a- a - a- I t s a-if a- a -’ ’ ’- - a-a- - a- ’ a - a - ma -ma- a- I i la - a - ‘a- _ i i a - r n a - a - a- a -t a - ma -a- ’ ’ a- 3 a a- Ira - a - a - a - aa-- a- a- - a - a - a - a - a -- a a- a- a- a a- - a - s a - a - a’- a- a- ma- a- ’ a-’- ’ - a - a - a -- a- a- - ‘a - a - -a - a - a - p a -’ c S a - ’ a - a - t a- ma - a -a- a-a-C t’ t a - a - a - a a - i _ i _ a - - ‘ a - a - i a - a - a - ta - a - a - a - ‘a- a- a -- a - - ‘ s t _ a- _ a - a - a -’ a- I f i a - ’a-r cs c-a -a- a - n e a-a -a-t S t a - a - a-a- a-a- c- ‘a - a - ‘ a - a - aa- - a- a- a- - ma- ‘a--a- a-’ - a - a - a -a- a- ‘a - a - - a - a - ’ ‘ a - a - na - a - na-s I a- a- a - I a- C a - ‘ a -f a- ’i ’ a- a - a - C t a-’ a- a- a- il a- a- a- a - ma - a - a - a - ’a -a-a - ’a- ’ i a- n ’a - U a - r l c a - ’c s a- ’ a-a-a- a-’ta - a - I C a -i a - a - a - a- ’a- ’ ma - ’ “a- a-m’ - — a- i ‘- a - a - m a- - -a - ’ a - a - ’ a- ’m- t’ r n a -’ ’a - a - a - a - a - a- a-a-a- a - a - I i a -a - a - i t a - a - a - T a - i a - a - s i a- ’ ’ a -a - a -a - a - a -m a - a - aflL a- - l t ’a-q j a- ’ ‘ -a- a-a - ma - a - a - a - a - a - a - a - t a- a-- a-a a- - ia - a -’ a- -a- a- ’ ‘a- c a- i a- t n a - - ia- a - ’ a- a- a- ‘ ‘‘ ‘ a- ’ ta-a-S a- S a -a -a - c a- -a- - s a- a- a- t ’- a - i ’ a - i a - a - a a - ”’ a - -ia - a - a- a- a- a - a -l i t — ‘ a-J S t ’ C a- m cm- a - a - a- I a - a a-- -a-a-a- - tO ‘ a- — a - I C’ a- a -” I ‘a - ’ a - a - I - a - a - ia - I a - a -’ ’ a- -a- a - a - a - a - a - a - a - r a - ’a - a - a- ’a- s ’ a- a - a - a -’ m- ’ a - a - a - a - a - ra - a - -a a - a - _ a - -a - ’ ml a- a - mi a- a - c a - a- a - c a - a - a - a - a - a - a - a - a - i a - a- a- I a- t a- a - a - a -t a - I a- - a - - a - a - ’ a- a a - a - a - a - a - - c -a - I a--a- a- ’ a - a- a - ’ ’ a- a - _ ‘ na - a - ia- ‘‘ ‘V a - i I’ ’ Ha - a- ia-a-T a- i a- ‘ ‘a - a - a - a - a - a - i a ’ - a-’ a- a - a - n a - a - ma - a- ‘ a- w a - a - a - a - a - n r a - a - v aa--a - c a - a- j a-a- - a - i ‘ a - a -I - a - a- - a - i- t a - a ’- a - a - _ ’a- a - i t a -a-a- r a - a - t a-a- ’ a- a- a- ta- ic’ i - - a a - U a - a- ia - a - a- a - a-- a -C a-C ’ ‘ ‘ a - a - - a a-- a- i a- a - a - a - a- - - - a-t i a - a - I r a - ’ a - t a a - a - ’ a- ’ a- a- a - V a - b -a I a- La - a- a - a - a - ’ a - ’ ’ a- a- a- ‘ a - a - a -’ ’ a -a - - a -a - i sa - a - ’ ”’ a- a- I ta- -a- a- a-” a - a - a - “ a - i a - -m a - a - a - a- ma - a - I’ ’ i a - H t a- a- ’ l ’ l I ’ I W ’ A ’ a - 1 L a-a- a- a - a - a - l a - a - n a --S t ta - a - - a - a- - i a - a - i a - a -- a - a - - a- - a - i a - i t a-a-a- a- a -a- a - a - i ’ ’ ci a- - a - a - a - a- a- ia - a - t a- a- a - a - a- a- a- a- a - ’ ra - a- a- a-I ’ a- a - I a - I t a - a’a- - j a- a - a- a - b a- a-a- i a- ’ f a- a-a-a- -’ - - a-I ’ a - ’ ’ a- ” ’ - a- ’ a- ir a - a -- a - s t a- m a - a -’a- a - I ’ a- a - ’ S a- a - t a - n C t ‘ a - aa - a -i - I’ a - a - a-a-a- a- a- ’ - a - a- - a - ta - a - a ’- a - a - a - O a -’ c a - a - - a- - a - t a - a- ’a- ’a- ’a- t a-’ ’ a- ‘2 a- ’ - ‘ s — a - i a- a - a - a - a- a - a - a - a - a - a a-- a- a - a - ’ a- - a- -a- a - a - a-a a--a- am’ - a-n l t - a - a - r a - t a’a-n- ‘ a - a -’ a - i- a - í a - a - a - c - a -n - a-C a- ”j a - ’ ’ ta- c a - a - a-a- a - a - a -a - a - a a- - ’ a - n ‘ C a -a - ’ ’ t ” a - m a-a - ’ a- ‘a- a - a - a - ‘ a - a - a- aa-- m’ i ’a - ”j a - a -a - a - ‘a - S a - a - a - a n- a- a-a - -a- ‘a- ma - a - a - i a -’ ra - a - a - ta - ’ a- ‘ 2 a - s a-a a - a-a- a- i a- -t a-a- ia - u a - a’ - a - a - a - a - HI a - n f a - a - ici’ a- ta- -” ’ a- a - a - ’ m’ m a - t i U a - C ‘a- ’ i a - - ’ a - ’ - - ’ I a- a -m a- ’ ia- a - ma-a ‘ a-’ a- a - -a - na - a - a- i a - ’ a - a - a - a - a-i a -’ L a - a- ’ a - a - t a- a- - a- a- aa- - a- ’ r - a- a - C a - L a - c a - a - a - i’ ma - 7 a - i a - - ’ t a - ’- a- a- Ca-’a - a - r a - a -a-ia-a-” a - a - a - a - a - taa - - f a - l a - a r-- ’ ‘ ‘ a -a - ’ a - a- a- ’ U a - a- - a-a- a - a - a - a - ta- -a- ’ i a - c a - a - ’a- i a - U- ”s a - ’ sa - aa -- ’ s a - ’ a- ’ i _ a -’I - a - a - a - a - ti t i a - a - a - a - a-a -i a-’ tm na - a - a - aa- - - a- Ia- ’ a - a - a - a - a -a - a - a - -a - a - ’a- a- ’ a - a - a - a -i a “ ‘ - ‘Ia - a - a - a ’- i a - i a - a-a-a- - aa- ‘a- - m a - m a - a- ’ a - a - a - f a - d a-- a- a - a - a - - a- a- a- a - n a - a ’- aa- a- a- f l a - a - a - a -a- a- - a - m i a - a- ’a- m ’a - a - fi a - ia - a -a- a- i ’ la-a- i a - a- ’a- ’ in a- a- ‘ ia - a - _i a - - a - a -a - ’ í’ a - a - ’ a- ’ ’ t t a - a ’- a-c a- a- - na-a - La - - a - fl ‘ a- a - f -a- a -a-a- c a - a- I ’ a-i a- a - a - a - a- - a- a - a - a - a - a - a- - a- a- a-a - ‘ a- a- a- a- a - a - a- a- s ’ ’ a- a - c r i u o f a- a- 1 L a - a - i a - a - U a -’ a - a- fl aa - a - I i t a -’ a- a- ’ a - a - a-Iou r a-i a - a - a -a - n ra - a -- f - 5- a- ’t ’ a - a - Si’i a- I a - a - v a - a - r S a -a at-y a- ’m a- a-a- ta-’ a- a-na- l a - t a - c a - a - a - i a - n a-a -a - a- v C m-a- a- it’ ‘ - a- ’ ’ to I a - i a - a - c a - a -- a - n a - a - a - a - a - ‘a - ia - i_ a-a- -a- a-a- i a - a - a-a - ’ ‘ a - f i a - a - a - a - a - i a - a - a-a-a - ‘a- im a- - a - i a - ta - a-’ a - n a- a-n a- a - aa - - a - a - a-a- a ‘ a- ’ i a- a- a - a -a - a a-- a - a - aa - - a - a - a-- a - b a - a - ’ a- a - a - I ’ s er - ia - a - a - a -- a a- ’-a-t c a-- ’ a- CX a - a - a - a -- a - a - a - aa- - i’ a - a - I Ia- ‘ a‘ ia-’ - aa - a - L a - ma - a - - t a-i a-a- - ’ a- — - a- I a - a - ” a- ” -a- 0 i’ — “ aa- _ -— ‘ - _‘ _ - a- ‘ — a- — ‘ a-- ‘- ‘--a- - a - a - —-- ’ l I a a - a - a-ti 0 u if l a- 4- 4J ‘a- UI 1 1 Q a-Q a - Q 0 -a-’-4 W CO N N Ci a-f - a--4 -a- ’4 ’a--’4 U U t a- f l s- icc w ua- cti 4i la--4 4 w a -c a a- iE- ’ a- ’4 a-’ 1 r1a - 4 ‘ “ I c t 0 O 0 0 0 Z O j a ‘a--l ’a-’I ’r1 O a- 4 1—i O Q a - ’i ca-j W a W ti’a - -a- - -4 E--4 E-4 E-l 11 4 a-a-’ a - - ” oiti ti o O ”’”” La- C’4 N a u U a-la- a- ’4- a- ca-I Q 1 4 W 0a -E C Q 0 a- • a-—4 cc a C 4a Q a -ti f 4 ta-i a-ti Ca - iW ‘ 0 — Z U 0 —4 -a-’I N C ta--4 - a- a -- 4-i 4 a ‘—4 11 4 o -4 0 0 z cc ‘ 0 la- -i eta-Iaa’ N 0 -a--a- - 0 i-a- N In 0 a- N m a- 1 a-ala--I 04 0 l QI W U — - 4 a- a-a- a- a- a - a - a- a- a- ’ a - ’ a- aa- - a - i a- a- a-ri N a- a- a- a- s” a - a - a - a - a - 2 a - a - ’a-v ’ a- a - a - a -’ a- ’a-a - ta - a- -a - a - a - a - a - a -C I 13 s i ze of W irt ía -a-a- 0 a - it a” a - a - a - a - T Ca - 1ia ’-a- a - l L f i q a - t a - a - a - I a - a A 1 - i a-a- s Oia-’ V C a -t a- a - a a--a- a- a - a - a pr ’eci i ct a -a- ’d for a-a- a - a - f l o r a - le v e l a- ’a-o st at a-a-If ‘a-i l ‘ a- a- a- a - a -a- c on a -a - e c ’t i on a-’ t a - a - a - a -a-a- p a - a - a - ta - a - C I ina-’a- - a - a - a -a - a I accompanied by a-a- hi a-a-a -leve l c o n s i d e rab l y by ic a-d t o t h e a- p a - a - a - aC - t t h e a-a-a-iX i a - a - us a - — ri_- a- r in sm i ssi a -n a - r-’ w a - a - - a - a - L a - ia- a - a a-- - a - i’ a-a- ’a- vi a - i m s a - 5K f or a-a- _a- c tri o US a- a - - a - a - a - a -a- n a-a- ’ ’ a- a - a - - r a - a- ” a- a - t y f ur a- a- n a -a - s I a - a - a - at- a -u r e a - of ta - a - t a -’a- a- ’ a- a-a-iit y a - a - a - ’ r a - - a - ”a- ’í a - t a - C C a - a - a - a i- ‘‘ a -a - a - t ci ’ na - a - ’ ‘a- a-t ii a - i a - a a- la - a -wa - a - i _ a - -a - a - a a-- s a - a - a -’ a - ca-’t i ta- ta -a_ a- a - i i s f a -a- ’ ta - l”v a-a-a- ny a- a- i a- a- a- a-a - i a- a- ta - a - a -’ a-a a - a -a - na - a - era - a - Ua- ’c r a-’a-a -a- a-at- a-on s s i on s i’ra -a- a- m”a- b oa-a-don l a - a - a -k is -’ L SDAC t’a - a - ia - a L1 t a - so ur a -a-• -a - a - 1a - cle a-it a-P a-a - up of a - a - l a -- s ‘ a - i a - a - A R P A N I’a- ’ a- I I P — a - 7 I P w e r a -a- o f a-a- a- a-a-c a - a - a - a - a - a - a - I l L — l e a - a - cl ta - a - 32 1% some ta-h ia - a - ’ ta ta - a - a - On a- iaa a - a - a -a- a - na- a- a - a- a- a s a tc l i i t - e a- t a - a-c’ WO a- f a - a - a- ’t Ha - rca- ’ - a ’- I a - P c ia -a-a - a-a -a-t i_ a-a- ’ a-a- a - a -a- a - a-a- v a-a-ira - t Ia-a- a- a- a-i 1 1 ’ ’ a- a - t a - a - l a- a- - a ‘ S ai a -’ a- - a- a- a - v a - a - a - a - I t a- a- ia - a- a-a-a a--a- ‘Ia - a- ’ ’ a - a - i li a - ’ a - i t U a -a- a- -f a-s o I a- a- a - a - a- a- ia - a - a- ’ ’a- a- I ‘ a-a -c a - a - a - a - p t C a - t i a - a - a - a -o f c” us a-a- a- T a - P a-ia-a-a-’b ’a - O 3 t a - aa-- a - i a - st ira-quish a-a- a- a-a- i a- -a - a-a - a - a - - a - a - a - -a - a - a-a - f a- a- a - a - a - L a - Ia- a-it a-a-c r a- t ra -a-i’ s r a- i s s i o n a - a - i a -ma m- a - ’S a- a- ‘a-t a - ’ — a- I a-a- a-1 a- a- a- ’ia- a- ca- ’a-a-’a - n e c ts t h e TCP ta -a-at k n o w i a - i a - ia - ’ a - i r a - ’a- ’a-l a - a - I u a ”i- ta-a- i ct ra- a-a- a- a-a-a- r e a l a-’ l i a -’ a - ’ a-’ a- ’ a - c h a -n a- a - a - r t a - a’u - l ar a-a- a - a - wa - a - a - c p r u t o a -’a- a- a - s a - a - a - a - na- t — ia -a-i hi a - a - ’a- ’Ca-i a -a-a -itY I a - a -a- a- a - — d e l a y a - a - a - a - l a - n a - a - a - a- ’S ’a - a - a-a-i t a-a-a-Ill k ’ i a - l SA a a- - a - ’ u a - a - a m a - ‘a-- ’ i i at a - a - a - a - a - a - s e a - a - a - c i a a -- ia a- - a - a - a - a - i a - i a i- ai -t a - a - a -a- a - c t a - i a - a- a - a - a - a - itt ir a - a - a-t a- a- a-a-- a- dy ‘‘ a - a - a - a- a - a - a_ j r Ia - Va - i occurs -‘i a- a - i t protocols i a - aa - I a - a - a - a -- a - a - a - a - t a - i la- a-a- b -a- ’ va-a- a - a - a - b a - a - ’a - a -a- n 01 ia- - a- i a - a - o w a- ’ c i a - a - ’sa - a- ’ nt Ha- a - l a - ‘ta -c m a- a-xia -T a1-a-1a-a- t h r c a - ’a- qhn a -_a-t tin ’’ i_ ha -a - a - - a - i r a - i a - a - t ia-a-a- a - c i a - ta - a - a ’- y A l - i a- a-A a - t a -a - a - a - s 1 Sa-a-a -R I i q u r e s sn o w m o r e t a - a - n i _ i a - i s h o w e v e r _ ‘a - a - i t ’ a- a- ’ I i - a - ma a -i’ a - a - a - a - a - a - ri a- ‘a- ‘ r a- ’t ‘a-a- n a - a- mm a- I I’tP r e n a - a - c o i l f u r t h e r a-a-a-’ a - a - r a - a - t a - a - a - a - I 1t ye I a-a -a -i ta-a-i’ ‘a- ca -w e d a- ’a - ’ t r a - I a - a - a - a - a - a - sa-S aj - n a - a - t ‘vi a - a - The side - a-a- 131P a- ’a- a - v a - l’ a- ’ ra-i’ ’ a- a - a - a a-- a’- a- a - l a - a - ia - a - a - a- -a- a- - -a - ta - a - ta - l a - c 34K na-a- a-a- a-a-t i o w — c a - i p a c i t y h ci h — d e l ’iy L a - a - a - a - a - a - g a - a - o uis t a- a - a - a- 2 I qa - ‘ i l a- a-a- a-- a- ’ — a - a’ - a-a-a- a- a -a - a - a - 8 1’ a -a - aa- -a - a - t a - a - a - a - a - na - a - a - a -a -va - ’ I a -tI ’ — i a - a- ’a- a - a - ’2 f i _ a - a - a’ - a- - a - f t a- ’a - 5 m a - a - a - s t —to— ia- v a - ’- ’ i n t r a -’a- du c a - a- a- a-a- of selsia -ic nata i a - ir e a - s u l f a - c a - e r a a-- a - a -’ A a - ca -a- - ’ a - r d a - n q i y 8 - 0 1 t a - n a -a- - Ia- a-i kSAj Of ta -a -j ’ a-a -a-Ca -C a-t 35 ca-’ a-a- a-a- ra- c’ i a- n n f a - a- ” a - a - t a - a - a -a-a- a-a- a-a- i a- a- Ca - a - ira - a-a -’ a-a- a-a- a-”- ‘l’ a-a-ca- a-a- f f a - c t a - v e c a-p a -a -city of to NORSAR 25 3 i a -a- K bps li a-’a - i t a - a - a - a -a- a-a- a- 1 5a - a - X -4 a-” a- a- ’a - i a -’ a-urtm -a - f l t C C a - a - ’t a - _ a- - Oi ’ r a-O a - a- ’ a- a-a I a - a -’ 4 l e t a - a- ’rs ta - a - is ‘a - a - fol a-a- w i n d o w of t a- a- Ca- a-a -s niora- ca - a - a - a - a -Ia-a- of 1 75 cetters per a- ca - a -a - a -rib a- a-eci a-a-ceu - a‘- na-a -a- a- a - f S e a - -I ior a- ur a -_’ ia - a-r e p a - a - ct a-a- wa- ta - ’a-e a- a-a - a - ’ i a- ’ a-ta- ’ la-a -a-a-e a- ’ r a t e of f i a -a - a -c a - s a- ha - v a - a-a-a-a- a- ’ -v a-’ a- a- a- in s a- a - a - a - l a - a - I a‘ a - a -a-c a - a - I ’a- ’ ‘ La - a - a -’ a- a- -a - imu p r of a - a - a - a - -“a- a-a-a- a- i_a - v a- aa-a- a-a- ia - a - a- ’a- ’r a t ‘ a-na - a-a- a - -a-m a- ’ a - l ea - a - a -w Iiea - a ‘a - l a a- -c— ia - -va- i m c a - s u a - a -- a - a-a-’ nt to a-a-a-a-a- ’ a-a- - a- I ia -a -- luV a -a- I n a - p - a - a- a - a-’ a-a- a - a - a - 1 1 ‘a-i t i a - a - a - a - s a-a-f protoco a- a- a- a - a - a - a-a- a- a - a --a-a-s la - a - 5 7K a- a -ps ra -a-t a - a - a -- r ta-a-ca-a-a- of a- - a- a- a - i a - a - a - ‘ — t a - a - — I a - a - i t ‘a- n a - ’O ta- a- ”C ’ a - a - a- a-a- a - I a - ’ ’ t a - _ a -’ l c a -a - t a -’ a - a - a - a - i a - i S w o u l d a-a-pa-a-a-a-a- a- to be ‘a - a - t C to ra -iama - a- a-- C a -a- a - - -a a- a - a - a- a- a- i c a-t e r s per a-c-cor a-d corr ca-poa - a- a- t o a n a -a-a- a- --—’ — ‘- c t a - ’ ’ a- - ’a - ‘ a- ’ t a - a - ’ — — ‘ ‘ ‘ a -a- _ a a- - a - a - a -- aI - - — a -’- —— - _ - a- i a- a- ’ ’aa- a - a’- ’ ’ a-a- t or t p a - a - c - a - c a - s ‘v i I ‘a a-a-a- -U 1 a - ’a- ’ ’’ ‘ a- a -a -s V a- a-a- a- ia - a- c a - a - a - a - — ’— _a - t a-a- a- S a-tm a- a-a- ’ - a- a - - ‘ ‘i ’CP a-’ - --i ‘ a - a - a - ’ ” — ’ 5 4 S i m u l at i o n 58 — Studies As a re nult of studies undertaken bu f f e r Cerf Pa-177 at BBN into strategy a num b er of propo sa ls were m ade re ga r d ing the r elationship of TCP window size and buffering strategy These w e r e i n v e s t igated at UCL by a s imulation study of methods of credit return In TCP credit is returned relative to the last a cknowled g ed packet and defines a con tiguous area e x t e n d in g in the data s t r e a m f r o m the end of the window t h e d a t a in the l as t packe t ack nowledged in and only in which the sender is allowed to transmit retransmit data Generally a s packets are acknowledged by a TCP on behalf of a rece i vin g process cred it is also amount of re t urned d ep e nd en t on t h e s t orage in t he Rece i ve B uffer Queue and in u nfilled TCP s t o r a g e a l l o c a t e d for the Receive Packet Queue app roac h es may be d is ti ngu ished Two general In conservative schemes the a-I located credit must be guaranteed by available storage which come from i nternal TCP buffer exclusively allocated ‘r ay p rocess a nd from any user buffers which are provided to the In the ‘a-- 3r iant proposed by BBN there are no buffers provided by t h e TCP ari d the window size is governed purel y by t h e buffers m ade -iva ilable for the connection with the received da t a b ein g placed d i r e c t ly i n tl’ - user In optimistic schemes of w h i c h buffers ta -a-c fixed control window used in the experiments discussed in Section 5 3 is an example some t i mes mo re c r e d i t may b e g r a n t e d i _ h a - a -n is actually available on the assumption that the extra si_ c- r a g e w i l l become available no later It is clear that provided su b net lo s s occurs a conservative scheme is 100% e f f i c i e n t since space is always guaranteed on the receive end vii th o p t i m i s t i c scheme s it is possible that be d i s c a r d e d as the promised packets E owever may have to space may not in f a c t be a v a i l a b l e when they arrive In considering these schemes we should c r L t e r i a by which they may be j u d g e d bear i n min d the ge n eral These f a l l i n t o two groups- resource protection achieving maximum efficiency minimum st o r - a - qra - u t i l i s a t i o n minimu m s a - - r v i c e u t i l ia - s a t i o n ‘i ’ r a - r O U 4h Cpu u t i l i s a t i o n by the TCP m a x i m i s i n g and throug h put m i n i r n i s i n g d e l a y p Ut will be governed by the minimum of the channe l band— w i d t n the sender production rate and the receive consumption rate a- a-a- a- a-L a - a- a- a - a- a-a-a- ‘ a-a- a- a- a- - a - ’ ‘ -— a- a- a-a-a -a- a- ‘ ‘ ‘ A la -a - i_ haa -a- -’ a- - ia -a- I i t t I a- a- a-a- r q a-a-’ ’ ’ca- a- a- va- a-a -n a- a - a - a - a - a - a - c a - a - a- e r ’a - a -a- a- - a- v aa-a- i a-a- soa-a-a-a-”a-a- a- ’i at a-a-a-” a- ‘a- rf a - i a - ’ a - a - a - a-a- a - a - a - t L a - a - a’- a - a - a - j eCi a - a - a - ci a - c o a - a - a - un t1a- I a- a- ’ a-La-a-a-id - a-a-- a - a - a- ta -a - ’n a - a-a- E _ ’a- a-a- a- - a-a-ia -Las ca - a - Ta - a- a- a-a - a- a - a - a - i a - a -a - i a - ” a-a- “ ' ‘J ’a - ’ it a- nr a-like y ia-s t a -a-a-a-a- ACa- h t a- a-s Tne - a - r i t a - a - r i o t a - a - a - a -a- a- a - n a- a-ne p r o t o c o l S ’a- a - i a-a- a- a - a - a - ’_ a-a -a - a-a-a- a-a- a-ia-pea-a- b succeed ia-a- optimising all fl a -a--vi control ’ cou sl y Ta-a- c s i m u l a t i o n is a l s o a- ’a - i f a - c t a - a - U by - ‘a- hj e c t i v e s a - i r a -ui _ an— will a-a- a- a- a -a -a - a- a- a -er of external of the kind we ha -a-va - notea-i a -ra - pra -a-’v a - a -’a - ua - s ec t i o n s t r a f f i c a’ a’ -a - a - ta- a-i a- r i a - en n a - a -tt e r i -a-s p r o c e ss a- a- a -a -a -i a -ct ’ i o ta - a-a - n ’a- a- ’ ‘a - a -a ’a- - a -a-a-n a-a- a- st r a - a - a - i a - a - a -io -‘ ‘ ‘va-a- r a - t s a - na - rc a r i a - a - ca - a - c a - aC- aw a - a - a - a- a - i a - a -a- e a - a - a -p o n c n t i a 1 in study v a - I r y l n a - suba - a - a - a -t a- q a - ’ a - a -a - a - a’ - a-i a- t a - a- ’ a-a- V o - a - a- a -n ” l a - a - l ’ 1 7 6 ta -a -c a- ’ a-r a- - a -- a - a - ’ at a - v e a- a- a - a - t a - s t i n t ’ a- a-a-a- a-a-ia- i w i t h a-s a- “ a - a - t i i ed a- ’a- a -” la - a- — a- ’a-a- r a - a - a - e r v a t a - V e BEb ‘11 c r s e a - a - a - a -n- a- a- ra - a - a -u i t s a -nd t h e a- a - a - a - a - a - t a -’- - a-5a -ia -a -a-a - r c a - a -a-a - i v a - ’ a - a - a - a - f l e a - a - a - na - c l Ir a - a -’ a l l fern ta-c a - i a - i a - wa-a-re compared w i t h th e of ‘ a - a - a - l a -a-a - i i’a - ’ ’ aia-a - i results va -a- r i a n t a- wa- a- a-a-” a- I ir a - a - a - Iar a - a - a - u i r ea- — a- a - a - a - a - aI- a - t a - a - aa- - a - e d a n a l y t a - i c a - a -i ‘i a- a- im i s t a - c a - c i a - a -a -ma -a- l a - a - vi a - a - ira - a - g e n e r al o p t i a - a - a - st Ic ta -ic S L a - a - a a- i L a - i t a - O f l H a - ’ a - a - a - a - V t o a - a-a’- a- c o n s e r v at i v e scia - a -a-a-a-a- a-a- ’ a - a - nd ta - a - a - a - a - c a - c t s su b a - a - a -a- a-a- a-a- n a-a- p a r t a - c u i a - a - r schemes ‘- ‘era- j o of sca-ia--a-a-e a -n activity hos t teta-’a- a - o r a r l i y in - a - I r I a - a a--- a-rr ivina - Receive One a- a- a- it a- ’f a- ’ a - a - nra-ia- r Q n a - a - ua- ’ P a- a-a- c a-’ a- a - t otha -a-r is b a s e d u po n t h e 13i1a- a- a - a - a - na - a - a- - aa-- - C a - a a- - a - a - ‘ ‘e a-v s a - n - a - t a ta-ca-n i_ s a - u a - l a - a- a- ’ a-a- is if us 01 t a - ua-a-a-’d — - I o c a - a - a- ‘a- a - a - a - SI ’ a - a -t a - a - - s i_ e r a - c a - a - a -or ’- i e n c a - t a-a- The Rec a - - a -a- B u f f er s made k nown to t h e s e n d e r This i l l ’ a - ’v a- ’a- ca-’a-a-a- t — o i — s e a -’iu en c e en try of a-a -ia-a-ta to Rca-a-cive i- u l f c r n w h e n t na - - a- c or r e c t b a - c a - a - v a - buffer f o r a-i pa a - t a - c u l a r p i e c e 0 1 a-’ a - a - t a- i i s rosa-a-’-a- a- re a- a - l t a - a - l a - a - I ta-e i w j t a - a - a - ta- n a - C v a - a - r a - o a - aa- a rc ’ disca-a-sseci a - a - a - a - ia - - t a i l ia- n a-’1 a- qa- I ’l 76 c a - a - a - i aa- - t a-a-c a- a-a- a-r ’a - o ’a-a-a - a - a- ra - qa - ’ ra -a- ’r a - a -i i a- a-a -u a - n a - I a - a - a - a- a-if I a- l a - a - U 7 r a - a - ’ a-a -umber of b u t ’ a - a -’ ra- - On ta -h I a - i a -’ a- t Ba - I a - It r e s u l t s are a - l a - a - a - a -i a - a - a l a - a - a--a- ir a- ’ a - a - i a - a - c ta- a - a - c r os a-a-It s a- - a- a - a - i a-a - a - - a-a- - a - a - ill s a-a-a-n a -emes here va-a - r ’ n a - - n s a - t a - y e a- a- “ a - -a- a- a- a - a - - a- - - r-a ’ -l ’ a- ’a - f l S L a - l a ia a - ta - a - a - ill n a - a - f l o r a - s r a -t a-a-’ a- I ia -a - a -n a-a- a-nra -li h a - a a-ia- -v a- ’ r ta - a - ra - ’a - a - a - a - i a- a - a-i a - inn a - a - a - Ia-a-v v a - a -v in u t i l i s a - i t i O t - a - a - a - a - a - a - a - a -’ a- a- - ’ ’n t a -a-a- a-a- a - a - i a - a - l a - ’ C a - n a - a - c a - p a - c a - t i ta -a- a - s a- a - c a - a - a - m e ia - a - V er a - ’ sea-si t a- ’ ’a- to way a- ’ b u f f e r aa - a- ia- a - a- la- a-l a-a-c’ a-a- n d e nt ‘a-a-a-a - i i’a-a- ‘vii I I a-roca-’a- a-’ ha-a- i i a - i vi a - a - ur la-’irqe ‘ a - a - a -’ a - l a - i a - a-I in t h r ough put a - - a - i ’ t i c u l a - ri La- I f igu r e s wi _ a - a - a- ’a- ’t u r a - a - i a - v a - t a - - o f ta -ca- BP- U a-a -j Za - a-a-a- ta -a-match da -” n c r a -’ i 1 sa- a- a-a--’ aof low a- a-a-a-a- a- a- a -a -t a - - ‘a-a-” l a- a-a- t a -a - a - a - - aa-a-- as a-c a -a -s a-ri a - i a - C n V L I a- ’-a- a - a- a- a- a - a - t t c r a - aa-a-- tT f e r a- a - a - a - a - a - Ca - iC is v i ta - a - a- ta- a- ca-a-a- ’ a-a-a- a - a- a-a- unrcl L a - a - r a - U a - hia -La - t ia - a - a - a - t a - a - a - a - a - C c i a -’i ’i a - ’ ’ i a - la - a - a - a - ia-a-a- ’ a- a- l a - a - i a - a - a’- a - a -a- a- a- a-a- a- a - a - a -’ a-a- -a - a- - ’a-” a- of a - I the sa-’ia-a- i 1 R e c e i v e a-a- _ _ _ _ a- — 6 0— 40000 T h 9 t No Receive B u f f e r s a- O0 L e t t e Delay Delivery m s 200 lOt 0 a- 1 0 2 3 4 5 ‘ a- ’ a - ria - a - r’ vj tj v e BUN V a - i r i a - i t i o n c-a-f Number of Receive B u f f e r s a- a- a- a- a- a- a - a - a- - ‘ - ‘ T a- T a - ’ a- ” ” a- ’ a - a- Scheme -- 5 7 ‘‘ la-i a- - 8 R e c e i v e B u f f e r size Letter size Mean propagat i on delay Subnet loss factor ia- na -i sis Simulation aa-c 7 Rece ive B u f f e r s No a- 6 400 bytes 400 bytes 100 m s 0 a- a- _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ó a- ci’ s a - o ru q e a - a - Va - a - a a-a- a - a - a - a - a’ - a - a - a - f’ a- rs ai _ b o a a-- rather b a - a -a - a- a - a-a- a- a- t few large a- ia-ia -a- fl a-ta - ta - a - a - - cost i i j t a - l f l OU storoqe q ca -n er a I c o n s or v a - it a- a-- a- ’ h a - ’ Sens i t a- a-e a-s SO a - a a--a-I V i t a-y p a - a - a - caa- -is ta- I t h e BB t a - a - n ra - Lfl a- a-a- I ia - ’ a- a - i a -’ 10a- a- ’l a- il ’ a- s e a - a - a -- ta-a-c sm a l l sa - za - a -h a - a - l a - ‘a - i qha - t a- a- rouqhput - a - c i a - a a-a-a-a-a - ca-f a - c a - t i a-a- a- a-na -to our a- a -a-a-a- a- a- a- ’a- - propagation delia - v a - a - e ta- a- s low these f a c t o r s a-re css c r u c i a l ‘ta- r I a- a-a-i ra- y a- a - a - a - a - i a - ’ t t U a-a -a- a- Recta-a-a- a-a- ye ca-f a- oa - a - a-’a- a-a- t ta- a-a- a- a - a -n a - S • a-r a-a-a-tch a-d to tne buffer size a-a-a- t a - a - a - a -- a - a ga -v a - a- ‘ a - ca-a- k — a- a-o Lc a-a- ’ ‘a - a -a -c a - a- ’L a - a -a - rs a- a - is of a - a - ia- int jt ja - S a - i _ a -a-t a - p a - s a- a-so apparent a-v a-ga-ia-Il amounts of a- t h e ‘i’CP P roca-sa-a-i na-a-a-ia-i’d t a - a - a - s e a-a- d BI3N scheme to achieve ta -a-ca- sara-a- t a - a - a - a - a - a - a - a - a - a - a - a - a - i aW- a h- e n b o t h scheme s L a -ta - i s t o r age f r o m t he a-a-a-a- a - d e a i i s e d w i t h or e r e s p e c t to p r o c e s s a- a- ’ a - ’a - i a - ” I OU L s t o r a g e R e c e i ve ii u r f e r s i n BBN a-nd Receive Pa-a--a-a-s t a-I a-a-a-a-a-’ s t o r a g e i n t h e General Conservative a-enema- s pa -n a - s m o s t of ta - a - a -’ t a - n e ‘ i n t a - c ‘fCP pe r forming t a - c ’ f u n c t i o n s a - a - a-a- ba -c- a- a - i ‘ ‘a - i ta - ’i - -a-a - a -I la - a- ’ ta- a-a ic ta- a - a - a - i a- j a - t i a - a - a - l ta - t ta- a - a - a - a- a-a-’ f f a - c i a -’n cy ‘i ’C i a- vera- a -a - c a- - a- t ora - a - e o v e r a- a- a- a - t a - ’ r ia- e a-s i a - a - a - a ’- - I a - a - c a - ia - a - a - a-a-t a - a - a - a- - a - a - a - a - l a —- a-a-t a-la-a-a-cd a - a - a a-c - a -a - ia - a - i to a- a-a- a - t I ‘a- r a - a- a - ’ a- a - a - a- a - a - a - a -a - a - a - n t ra -a-a-’a- a a-I i f l ta - a - a - a - e r of the Ot a - a - a -’a - s cta - a - -a - a - a’s a- - i a - a -f a- va -a -n a - a - a - a -S a - a- a - t isfy ta-ha-’ f’a- a - a - w coo t r o l c r i t a - a- ’ r a - a-a- 11 a - I ’ ‘ ‘ a- a- a- - a - a - c a -a - a- na-c ‘ a- ’ a - ia - - a- a- - aa-a-a- I a- a - a - a - a - a- a- ‘a - a - a- a-’ a- a- a2 a- 0 2 2 a- 1 a- - a- a- ’ r a-a-a- a- if a-’ a - a - a- a- a- ’ i r a- a- a- a- ‘ a- r a- aa - a - a - a - a - a - I a -i a - a - a - a - a - a - ’r a- a- a- a- ‘ a- a-a- - a- nra- a- a- a- a-t a- a- aa - a -a- ’ na-c a -i a - I C Ia -- a - a - a -’ a - “ c a - a - a- a- a- a - a a- - a - a - a - -a -a - c a - ” ‘a- I i a -a- a - a - a - a - ca - ‘a- 1 — ‘ -a- - i a - a - a - a - Va-a- I ‘ a- ” a- “‘a- a -a - Pi aa- f a- a - a ’ - 0t “ a- l a - a”- a-a- a- a- ia - a - v a - a - ’ ‘ ‘ a - a - t a a-- a - n a - t a - a - a - i a -’ a -a- ’ a - a - a - ai- t ’ a- a- 1 it a- a- a- r a- 1 a- - i Qa-S a- a - a - a- - ‘- a- a - a ’ a- 0 a- t a - a - a - a - a - a - a - a -n a - a - a - t a - a a -- a - a-a- a- ’ a - a - c a- a - a- ” a-a- ‘ a- a- a- C S S a - a -a- ‘V a -a - ca - a - V a - a- ’ a- a a -t a a s i a - i a -- a - a -a - i a-I U C i a - a - a-na- a- a- - ‘ W - - i ‘ a- ‘a- ‘ a-a-c ‘ a - a - ia - a - I a- ’ - - S a-a- • ‘i a- ’ a - a - a - _ a - aa-’a- ’ ‘ ‘a - a-a- ’r a -a- l ‘ aa - a - a -a-a- c a-’ ’ a - I a- a-a - B ‘a - a - a - a- a- - ‘a - a - a a- - a - a - - a - ’ ’ a- ’ ‘ s a - - ” ' r a - a - ca - a - a- a - a - a ”- ta -a- n a - a a--a- a - a - a - a- a-b ‘a - a - l L a-’a- a-’i a - a- aa - l a - i a - n a - u n a-a- a-a- i i” i a - T u r e 5 8 a - r c a -a - f l y sa - a- a-’a-a- a - a - a - a - isa - s La - a - a a-- e x t a- ’ a -t a- —’a- a- a-S a- c a - a - a - a - a - a - ’a a -’ a- a-i a- a- a-a-a-a- ‘ a - - ”ra- a - a - t - a-a-” a-ia-I a - a -a ’ - ’ a - a - - a- a- a- t a -’ a-a- • a - t a- a- a- a-s t a-a - a - a - a - a - a -Iav - l a - a r - a-a- a - na - a - a - t a cr - a-a-” a - a - a- a- - a - Va -’ 1 ’h’I a - t a- ’ V ’ ’ ’a _ _ _ _ _ a- ’ a- - - a- I is seen a- a - - tha t no single sch e me m a n a g e s ta -a- a-a-i a - n a - i a - y e a l l flow ccn tro L objectives Some are achievable only at tba-e expense of ’ oa - a-a -a-” r’a- or throug h special pro cess behaviour In practice choice Oa - a- particular scheme for a comm u n i ca t ions sy st e m will depend on the priority attached to each objective This in turn will depend on the availability of resour ces wi thin TCP using hosts the kinds of communication envisaged and whether or not processes served b y TCP ’s can be expected to participate f low control objectives The use of a particular scheme by a receiver transparent in achieving TCP to the send TCP with respect to protocol is completel y fura-ct ions and interferes very little with protocol functio a - ts at the receiver Thus communication between machines using different scheme s is not dis b arr e d and e a ch TCP in a ne t work m ay use one several or a l l t he ab ove schemes depending on resource l im i t a t ions and its requirements 5 5 Conc lu sior a- The lessons TCP fall t ha t can b e drawn fro m the UC L exp e rienc e wi t h t he into two broad categories those that related to our s p e c i f i c implemen ta t i o n and those of more general inte rest Man y of the implementation lessons we learnt we re not drawn directly from our experiments but from the background work in settin g them T C Ps w as large and complex typically t he size of t he TCP w a s of t h e o rder 8- 10 K It took ove r 6 months to i m p l eme n t and was up The UCL TCP in co mmon wi t h mo s t of t he o t he r early a ccompanied by long and paira -ful debugging sessions c i r c u m s tan c e s i t is not s u r p r i s i n g t h a t our local In these meas u rement s show lo w th rough p ut a nd hig h processo r time The onl y first genera t ion TCP that did not show these de fects wa-s the TCPO a sin g lec o nne ct io n TCP implemented on an LS I 11 at S R I whi c h was re nde red much simpler by discarding the multiplexing function Obviousl y th is is not in general an acceptable choice but a- d oes it r a ise t he ques t ion of what protocol simp l i fi c a t ion s can r ea sona b ly b e made which we shall return to la ter Lat er versions of TCPO have included l ittle L increase the multiplexing in program size “ f a- nctio n with relativel y “ - part o a-no a-a- a -n d O U D t e d l y a- a -j a - a - L a - a - T ’ C t c C a W- a - a - a a-- a- va-a- a-a- a- a- c a - a - a- -iia - n ova-a-nra-cad costs a- a - a - a - T O t a- f a- na -it ca -s of w a-a-a-ste a -a-enti fied a-a-i s soa- a-a- ‘ “CX a-a - S ’f S a-’ r a p s were ‘ a- P I a - a - ta - a - a - a- a-ve rhea -a - a -a - i s a -a -a -nd the r u n — t o — c o m p l e t i o n ‘a - a - I a- ira - ta-a-c Ba - a -N as a m a j o r sou r cc a-a- ta- a -a - i a - a - a -a- a-ci Sa - a - ‘ ‘ ’ ta-’r i a- a- a-r e a- a - a- -a-ia-a -h a- i -i a- a - a - n a - a - a ta ‘a -a- r a-i t a-i ’aa- a- a- ‘PCI’ s na -a-a- b a - i a - a - w r i t t c r a - ra- a-a-a- the first l y o nl y implementations a- -Ca - a - t h e o r i gi n a l Ca- not attemot a- a - a - o u r own a - r a - a - a - l o m a - ia - a - fl a-a- a - U I L a - a i- a - i a-i l i st ta - a - a - a - a - I a- a - t a-a- of ia - a - a - a-a-a- a- a sp a-a -c a - f c c a t - — d e l i a -v s ‘-‘ a- a - r c lea r n t l e a - a - a-a -a - c a-’- ‘ a- ’CP iS i a r a - - a - ’a - a - - a -a- a -a- a - a -a -v a - I S a - a - a -r a - f ora -b ‘ s - ‘c aa- - a- a- c a - I i s a - a - a-’ia - a - a-J a a - a - i t a - a - a - i a - ’ ia - i a-a- ’fl ia- n a su c’ a-a- a- a - a - n a-a -a - s a a- a- a- - a - V a -’i a - ca - a - f o r I a -t a-a-r e a-a- m- ’ a -ne ‘a - a - V a - a - a -ta -ia -a -aa- la-a-c I a- implement— a-na -a-a -I a- a- a - a - a - a - b W IS w r i t t ’ er a - a - s p e c i f i c a t i on ‘Ia-h ere is a-a- clear need for ca- n-’a -ent at ions a- to a-a- ’ na - a - a - a a- - l a- I n c o n a -r a - l e t e B C P I a ta--ia- wr’ it e t a- a- ’ ‘ a-a -- a-ma-re funci ara-e a-a- t a-a- l ‘ a - a -’a - a - a - l l s s i a -’a-a - a- ’a- a- a-a-2 ’a-a-a- t a - a - a -- comp letely sa-eci t i a - a - a- c e a - i a-a- a-oug h we did have a n a-a-a- a- a - a - a a- - a -n a - i a-a -t a -c r a - a - a -a-a- TCP a- a- ‘ or ia -a- stanc e we ‘ a- ’ a- ’ I ia -a -aa - a - c h e cja - a - i e r — 10 the ea-i r ’l y a - la - a - aa-- a-a- l a - i a - ca - a - Ofl th aia - a- a- i t a - f l t r a - a - a - 10—fold -a-a- in i a - r a - p r a - a - v a’a -- a - a - a”- a - ’a - t a- a - x a - a - ”a - r i c ’nce of la-a - a - k a - a - f a - a - a a-- rotoCOtS of t a - ”a- S orde r a -I complexity “a-l W t a - i a-a-it ra-wr a- t ira- a-i na -cr c 50 1 i n c a - a - a -I c a - d aa- t a - a-a- t a- a - a - ca - a-- ”r a -t jna - s v s a - a ”- ma- a- a -f Howeve r r e - a - s o n f o r t h e p r o b l e m i s s ea-a-p l y w nca- ic a - a - a -n c v aa-a- a - a - - a -a a- -a - a-a- a-’ a- a-a- a-a- i a - a - h a-t ’’e ia-ia- ra-a-a-ia-a-a- a- a-r - a - a - on a - a -a- ’ - ‘a - a - a Ga- ’ La- a - a - a - a -lcp a - a - ’ a ’- nt a - L a - c ia - a - r’a-a-a-’a-a-i a - a - Li 1 1 a- a-a-a- S aa-a-a- a - ’ ra- - a- ’l a - ’ a- r i y f a - a - a - r a - y a - a - a - a -a- - a - -a - a- a - a - a- cia - fl BON I nt ena - a- ‘a-h ta - a- cia -m t ta - a - a - ‘ ia- a - ” t a-icy a-t p a - ’ a - a - a - a - a - a - a a- a- - al - ta - a - a a-- a-a- a - i k e a-f v i e w a - a the ta-a- w a - a - a t aa - a - n aa-- n a- a - a -a -a - a-n d Ia - ow w i l l a- a - a - a - p n o L a - a - a -- ’ l t a - a - a - a - t ’ a - f l O f l t a - t i a- f l ‘ a- a- a - a - a ’- a-a-a-S a - a - a - a - ca - a - a- a- a’I t a- a- p a - a - a - t a - i- -a - a - I — t a-a-’a- a - a - a -’ a - a - a - a -ra’i- t a- ’ a-a- fl t ro a- Ia- nd a - c ’ a - a - ’ i a- a- - a - aa-- I a- a - a -n t a - a - a -a- a - I a -’a- a- c a - a - a- a - a - a - c - a - l a - ia-a- a-” a - a - I a-Cia - a-a -icy a- a - a - c a s lm u ’ta - a -t a- a-ca-n i-a-” e- rk a-a-i a-a-a-’ l a - a - a - p a - a - a -a- a - ta n a -p i a- a-na -a a- -a - it i t ci a- a -a - I ta -T e a - a - 5 i r e a- a - a - n a - - a- “ a - a - a - a - i” a a - t a - a - na - a ’- r i r - a - a - a- ’a -” a- u a - a - a - a - c e a -ra- ai- l a- a - a - a - a -a - s c a - i n a - a - aa-- s a- ’a- ’r’a - i 1 a - a - a- a - r t i a J I v a- a- a- a- a-a- a- - n a - - a- a- a- ’ cx a -a - a -’’ a- a - a - a-a -a- a - a -’ a- a- a - a ”- a -a -nd ta - a - a’- ‘ a - a - a - d c ’- ’ o f a - a - a - -a - ‘ v a - i i l a - ia’- a -l a - a - b u f f e a - a - a - a - a‘a-a-a-e s a-spec a- a - c a l a -v a descrip tion From t he a - a - a - a - a a-a-a-a-a-ata - ti a- r a- a- receivers era -ni -a - a - n a-a- a - ta - ca - a - a - a - a - c i a - n t oa- V a - a ’a-’ - a-’ a- a - c ’ - s ‘ a - a - a - a -ar’- a -j a ” ta- a- a- a ”- re ’ Ly t a - ’a- a- a - ‘a- r t ‘ “ a - a -t a - t a- r u a - a -y ta - a -c correct posit ion a- a- a- no a- j a- lo ng e r “ c a - a - i a - a - nt a - a - f l i r ’a - a ’- J ” a-a- a - na -’ ’ ma- a-a- n d r ’a-r”a- a - a-t a - a - a - process of a - Otter - I a - a - L a - a - a’- t a - a - a - a - a - ta - a - a - a- n into a- a-ia- are ‘i a - a - a - - e f fe c t a - s ra- ’d u ced a-a- ’ C j a -- j a - n S ta - a - a - a - ra- ’ a - a - ’ i v i n q a-i ia - a - er process a - a - a - i t he f r a - a -a- n ‘i ’CP h ei ra -q t o be a- flq process ia-fl ‘- a - a - a - a - aa- - ra- ’c ia - a - a a- - I a - v e a - d a - a - aa- ’a-- ‘ -a-r t h a - s a- a -a - a -a - a - s a - i a - a - a - la - a- a - a - a- a-a- a - ta - a - t a - va - n ir t c a - rr i er a-a-lac ed a - t a - a-wa -a- a - a - t e a - n a a - t e a - a - a - a - a - -a - a c u r r a -”a - a - a- I a -a- a- i s a -a-I a-a -n a- i a-a- a-a’l- f l i a - I a- a- a - a - a - a - s Ca-a-a- a- a-a-a-- a - a - a - a - f f aa-cs - th us a - a - sc r r c a - a - a - a a- - a - a - a - a - a - a - ia - - na - a - a - la - ta - a-- ta - a - a - a - a - a-n aa - a - a - a- a-a- a - a - i t a- Ca- ’a - i -a - t a - l l a - a --l a -a-n a- a - a-” i t a-a- a - a -a - a - a -’ a - a - a -ta- a --a- ’ a- a - a - ’- ’ i a - a - a - a - n a- a- a - a -na - 1Va - i ra- a - a - a - a - a - -a - i a - a - a - aa -- I a - i’a-a- ’ a- i a - a - a - ai- a- e t ’ f o ct i v a -” l a - ’ I - a- ’ --a- -a - ’ - - ‘— “ a - _ _ _ _ _ _ _ a - - ’ As the simulation studies noted t h i s ra-duct ua-a- in intelligence a- - a- kos the throughput delay chara cLa -ristics a- a - a- u a-”i a- ra-- a - a - r - a -e nsit ive to In p a r t ica -a - l a r e f f i c i e n t ta - a - a”- a-eh aviour of inu iv ia -dual processes a-a- a- i r a - a -u a h p u t requires that s e n d — l e t t e r size be mu tchc a-’a-l c l o s e l y to arca-a-a - a - v a -—b uffe r size and the end-of-letter f l a g t hus b ecomes an indication to the sending process that it should pad out the rest of the current transmission buffer In other words the concept of “letter ” takes on the specif ic na-eanira -a- to the TCP of ‘a- ra-ce jve buffe r s i z e ” Since the e a - a - r l m e r f u n c t i o n of “end-ofb e t t e r ” as initiatin g reassembly automatica -a -lly di sa - i a-pca -a-red in BBN scheme t h e idea wo u ld a p p e a r to h a v e a-a-ta -a-‘a-fl a-fiG a- a- u t h e To a user process a letter ta- h e rotc a -col ia-n the proposed revision a-s intended to be a unit of data with a--a-ci t a l l no logical sia - a - n i fi c - a - n c a - u s e r processes w i l l h a v e need f o r such str ua - - t ur lra -a-3 and a-a-a-a-ose s that do will contain indications of it in t h e data field ia - a - a n y case so the user—orientated definition of a-a- letter would a-a-iso appear to be redundant At least under ta - a - a -’ BFN scheme ra- a-a- ere fore there is a strong argument for a-a-” i t n t i n a t i n a - a - t h e “ l e t ta-er ” ‘i ltoa -jeth er If t h e simulation studies suggest t a - a - a - it l i l a- potent iall y simplify the relation ba - tw a - c a -a - a - ta - a - c processes the crossnet e x p e r i m e n t s TVP to relate to existing ARPANET a - a - a - aa- a-e - ar to ba-a- redundant for p o i n t e d out Pou zin a -n a - i between protocol t h e u ser f i a - l u r e of t h e structures pa - a - a - ta -a-a-ca - -i - This would the potential layers has al ready been f o r t h e X25 p r o t o c o l a g r e e d f o r PT ’r n e t w o r k a ccess 1976 We certainl y intend to study the implications of importance of these interactions ca- a- a -a - a -a -u nia - a -atio nis to be hand led by activit y in 1977 depends very much on future direction of inta-erneta-working ta- a-a- a-a-- ‘ta- a - I a- in d i c a t e t h e this situation as part of our internetworking ‘Ia-he p a general problem for intern etworki r ’a-a - i activity a-a-a-a - a - a - a - a - a - I s a - a -I a -N If all intcrnetwork a- network of qa - a -tc wa vs a- a -ntinq a- transit net then we have a very clean design environment in ‘i ii a - i ch t h e p r o b l en a - d i s a p p ear s 01 - ‘ra - d na- ” r ca -a -ra -a- ” ’ita - a-natcd On t h e o t l a - a - a-r as la-and t h e situation ne two rks across which we m u s t m a i n t a i n an e n d — t o — a-a- a-a -nnec tion of a-n i n d e f i n i t e n u m b e r of n e t w o r k hops p r e s e n t s ioa - ia - a - d i f f i c u l t i e s a - f t h i s k i n d a - a - a - ia - a -ti t a-a -a-c ars to he needed is a-i for a -ny e n d — t o — e n d p r o t o c o l d e f i n e d “ network i n t e r f a c e ” w h i c h w o u l d a- n a - ia - d c the protocol to decide when certain functions should ta - a - sunj a-enc a-ed a - r a - a - a-at ’a-ols L a a- T h i s was done on an ad-hoc b a - i s i s to the ARPANET governing the establishment of a v i r t u a l a- a- a- a- - a - a a-- connection since a- — a- a- a- ’ a - ’ a-a- a - a ’- a- a-a- a- p r a - a - C O a - a a-- a-a - r a-’ a- f a-a- va-a- a- a- a - a - a -W a-’ Va- ’r - n of a-C a- ” a - a - a -- a-a- a- a- a- d a-a-a-a- a-b i n a-a-a ’ a - “a-a-a-ia- a- a- ‘a - ay user a- ” a- a- _ a - a - a-a- a- a- —a-c a - a - a - a - waa-i - ” £ £ poss a - o le ” ra- ‘i’CP a - a - a - a-a - I a - a - a- a- L a - a - o aaa- a - a - a - na - a - a - a - a -- a-inK a- a- a- a- 1 in k ‘a - w a- a-a-a - a- ’ The 0 was suspenc a-ia-’a- a - a - a - reed in advance a-a-a- a -a - a - ta-a-il to suspe ra-d a u i k t a - t and a mo re forma l procedure should be a- ”l a - ’- i r a - y net a-S a- ’ ‘ ‘a-a-a-’a-’a- — — ii’ I a - i a - a - a - aa-a-a- In - a- a-a-a- a - ta - a - a - a - a - a - r t a- ia -a-aa - la - a ’- a- a- a- ra - a -fu l ‘‘ ‘a- ’ a-a-a- a -a - i a -3 fl La - a - a - t aia- a-a- a - a - Ak a - a- na -— t a - a - — a-a-nd a- a -Ia - ta -rC a- i a-ua -C LUS IOa - i S we have d r a w n a - a - a - a - -- a - a-- st c a - a - a - a - a - a - a- c a - ’ wa - a - a - c a- aa - a - a- a- ” a - a -r ot a - a -a na -j resa-ons a - a - a - a lity - i a -i aa - i a-a-a- i rem a- t - a - a a-- a- a-a-I t h a -a- a - i a -a- a- a- a- a- a 1a- a - ct p a-a -rti a- nec a -IS t o a -n a - a -i a - a -i’a- a-to col a - a - a - ta - a - ia - a - a - a - l a - aa- - ’ a - t a - a - J a -’ a- i t a - a - a - Ia- a-a-a- a - u a - d a - a -a-a- a-fl a - a - a - ct a- a - a - des igne u s m a l l e r a- a-a -ore a -nc oL4a - i t C ’rs i n t r aa - a -a - a - it to c ’ f f i c a - a - a ”- a - a - a- a - a - a -d i a-i - -aa - a - a - au - a- ’ss a- a- As a ca - i rr i - a -r f o r a - i a -a- I i f l a -i a- ta - ’a- a-c’ a- a - a-a-i for protoco a- f aa - a - -c t ion a - a - a - a - ct ‘a-a-a- - a - a - a - a - a - i a - s e l f these le a ds uric shou ‘i C C it a - a -n g a-a-a-ia -’ ’ 15 1 a - ’ a - ia-S a - c a - s r a - t a - a - c r t a - a - a n m a - a -a -’ v a r i o u s n e t w o rk p r o t o c o l s ‘ a- ic a - a - a - n a - f i t f rom more it a- a- ou a- a - i a -ia- ’ a - i L J f l a - a - a - i a - e l ’ ta - a - a - n u s a - c c ’ a - s a - f l a - a- a- ’rna-a-”tWor ki o n i t i a - ’a -t n o t f rca-a-a- o u r e x p e r i e n c e a- a - a - i a - - a-a - h As flow an a -a - f d i v a - d i ng a-a- ta- a-a-i By the f o l l o w i ri ca- i a - a - a - a - is de tter C a - ’ i ’ a - ’ a - l C a - t a - a i a- t h e ex a - st— a- CIJAP TER a-a-’ l UCL A C T I V I T I L S t L a - S S la - T lj Introduction in Section 4 5 Kirstein 1 97b A we me nt a -oned our p l an s ca- a- a a- ‘ a - a - ’a-oc t L i ’S S one PDP9 to the UK P0 Experimental Packet Switched Service The UCL c o n f i g u r a t i o n ha -s been shown in Fig 2 1 a - i r a - a - g a - a - n s a h a-i s been slower than antici pated mainly because some unexpected delays ha-a-u to Our EPSS h as a-t the beginning of 1976 EPSS was provi d ing an echo service for a- ’ a - a - a - a - a - a - a - w are testing by which all packets sent were sca -bsea - u e nt i y returned Ta-a-a- service was gradually advanced to provide calls to oneself c a l l s te t a - s t number and calls to other users on the s a m e Packet Switching PSU L’ n i t This last state lasted a considerable t a - a - a - c and was made a -a - a -ore ciif ficul t by the availability of only three 48K bps ports on ta - a - c London PSU to be shared between the five users rea -tuirin ca - access ‘ is Monday to Friday mornings onl y 8 45 a - a - till a- a-a -ta- ia-a-a-I rOVem C nt in t h e level of service wa - a - a - a - a -trunk situation a - i a - a - a - a - O Wduring ia- a-a-ic a - a - a a--a• a-a-rime c a - a -ta -ne in time With minor improvements ta-a-is is the A PSU became available in Manchester a-nd in to modify our communication hardware to cope with had ’a-a- modifications are discussed in Section 6 2 V a - la - a - t h e lower l e v e l s o f t w a r e in a s t r u c t u r e d way CO a- a- a -cr c ’ i s a-ia -cussed in Section 6 3 e a rly ’t a-J a-a- it a - a - a - ’ a- Uc a - -ca- a- a - ri to a - a -- a - ’ a-- a - ia - a - i - a - a -a - t a - a-a-ía- a-a-a-a-aI ri -i a - ta-a - h a - a - rt a-a-a -a-t adapted leased Our p r o g r e s s t a - a - e r a - f o r e both ta - a - c R u t herfo r d o ur s elves decided to m o u n t an e x p e r i me n t a - i a - a - ply We decided I t was f e l t v er y i a - a - a - p o r t a n t to rea l e x p e r i e n c e w i t h EPSS uaa -a--a - r a a- -tor ’ a n d a- ’ A ti-a-c l a t e Sura-a-ra-a-er the course of 1976 a- a- a- a - ’ and this switching a second London PSU a - a -n a - I acce ss f r o m c h a r a c t e r ta - a-a-a- a-a- ir u i s became available a-a- a-a- a-a- a- a - a -n t 13 00 to i -a - - t a - a - a - a - a --u a - sa- to what i s n o r m a l l y rc a -a - a - ira - led as a-’ a - a - a - ’ J ‘ a - ’a -i a- Sc’ia-vice was and a - la - a - ia - fl we and the Rutherford Laboratory are two a- a - f the software module d r i v i n g line to be able to make a-i ca-ill In t h i s ta - a -a - co na -ra -a -u ni cation t a - a - r o u g h EP SS ra - a - a - a - ny more modifications since t h e y drive EPSS via - a front—end This activity is discussed services better via EPSS in Section a - 4 We still plan to our immediate p lans a-re discussed a- a- a- a-a- a- a - a-aa - -a - a - RL ijrdwa-a-ire and In ar a -a - e r to connect Line Handlers to EPSS we purchased a Transmission Protocol - _ _ - 67ai ts ‘a- ’ a - a - te r a- a- a- L ‘ a- a - a - a - v c a - O a - a - ma- ’c t ’ a-a -a a- COSt blect ia-a- c a- a- aa-- ’ a - j a - ia - a - f l C C a-a-a-sses a - a - r a - aCL - a - a- 5a-a- a -r a - a - a - i a - a - ca ’- - a - a - a u- C hC the and ‘a- a - c a- a-nd t he a- a-- er or r a - a - a- L i a -n a - c the on L uoa-irds can be Ta- Sa -a- c a -a- a- neces s a r y Ia-a-a-a-c i f l it a -u l i s at i o n ta-a- t a --a-u loop ara- a-a-i a- of c a-a-n a - ’ ’ a-a - ir a- sea - a - a - ca - a - ce uc g e n e r a t ed t a - c a- i’ a - a - c e s s o r so t a - a-a- a-a- ta -a - e p r o c e s s o r usea-a-i a - fp U in c a - a - a - i d a - a-aa- t r a n sr ’ i s s i c a - a - ta -a-c On ta - a - c V24 interfa -a -ce ar c s i g n a l st a t u s f r o m t h e to a-a-d t h e Ca - wa-s a-a-a -a-a -u‘ cta -u ra -c a - ia - a - a - t I -b w a -a - i a - a - i a - ‘i a - P U ta - a - c t a -i has a V 24 a-a-a-a-dora- i nt e r C i a -U c o n t r o l sia - Ja-a-a is of ‘ a- ’ u a - a- a- ta- a - a - a - a- r e t r a n s a - a - a - a - s i o ns 3fl st - a - n uard i n t e r f a c e t a -a - it Va- ’ i a- a- SO a -a -a -at _ jt passes a-ia- i packets received to ta-a- Sa-a-flu a- a- i a- a - a - c a - aa- -a- L a - n users a- a-5r O a- of s t at ua- b y t e to ta - a - a - ’ c a -U a-a- _ The TI’ U checks ia -a-cur ing p u c k e t s f o r CI K C a-a-ia-a-kut ‘a - ’ ’ - a - a - ia- L a - t a - a - a - a - a - Ia - £ a-i a-i a - O a- i’Z a - a-a-a -a- a - a - a - a - a - a - csL t d f l u na -ja-i c a- _ CCL a- JYa -a -u w a - W a -a- a- a - a - a -a- ’ a - a - au a - a - p p a - a - l u s a- _ - - a-a-a-ed a- to c o n t r o l ta -a-c T a -a - c T P C p r a-a- ’ce s s o r rra-easurea-’ a- ’a -ent proced- l ay h a-a-a-a- me u r e s r ca - a - a - a - a - red h i’ i’a-PSS w i t h o u t p r O c e s s a - a - a -r i n t e r v e n t i o n s ‘a- _ a - Ia -’ a - a - I a - a -a-a - a - i a - a - a-a-I a - a - a - a - a - a‘ia-- a -a W a - ta - a ia -a a- -u a- ’K a-a- a- I ta - a - a ’- syn c h a- ’a - n a -a - ’ a - a- Ohi a - ’ t h e Ca - C ia - a -u to a-a- a- a-a- cd - a - i’ L a - a -a- ’ ha-ha- l a ‘ a-’ a-CC ’j a- i a - L a - a - l aa -- i a - a - a - a - t a - c a - a‘ a -- a - a - a - a - a - a - a -i t ta -u C a - a - a - a - a - a - ca - a- cc a - v a - a- ” ia - a- a - l i a - La - a - t o for t a - a - a - s so ‘ i a -a - c a - a - y a - a - a - a- a-a-a- in a - a - a - a - sc a-a - Ia - c h a -a-f tW a - a - i a - a - K t ’ I a-h a -a - a- a - a - - i a - a - t er ra -- a - i ‘a - a - c ’ ia- n e s i r a -’a -p i a - l i a-a-h - a-a - a -a- cCa- ’N ” wa- a - a l a - i to a- a- not he coda-n a- a-a- t wa-a- I ab a-a-a- ta- a- a- c a -a - a - a-ic a-a -c ‘a -a-a-a-ed 6 5 wa - ca-a- ta- he TPU is a -va - a - 1 -’ ia -as a a-a -a -ca -a - a-a- i a-a -Ca a-a- a - l c a - a - u a t e ly a aa- a - s a- ’ “ a - c ac k ’aa - i a- a - t h e s e ea-a-cr a-— a - a - c i a - i a - e r of a- an if p t a - a - t a - cd written ‘ a - i a - “ a- ’a - a - t a- The s o f t w a r e It t a-a- a-a -a -I a- its lost a - a - a - a - cr c a - a- a -c t o be a -t a - t o tO seca- a-a-d Ca-a-a-a-C ta- a- a- a- a- a- i a - a - a - a -a a- - t o r t a- Wc V a -a -r c a - a - h a a-a-- h a -a - c t i o r a - ta - a - a - a - a-i da- ’a -p t o r a n d to t h e a-fPU t o test a --a-a-a- a- arrives it Several La - a - a - a - a - a - ia - a - a - a - t a- a- c Tia- L’ wa-a- - re t e s t a - a - d us a - oa - a -a - a - f a p o i n t a - a- -cil i a - a a--’ Ca -’ a - r a - o a - a - uty i u ia -C OV a - ’L’a-’a-a- a-a-a-1 r t i c a - a - a - r i ’ ’ a -C a JV a a - n a - a - a - a - nii a - to a a-r ca - a-a-a -a-s w a - a - a - a - a - a -a a- -a - n a - n aa -- t a - ’ a - a - t a - t a - a - aa - a - g a - a - a - a - ’ a-a-a-a- ta -a -ps a - a - i a - a - t a - i r Th i s Va -la - a- a - n t h e PDP9 ar a-u t h e c a - a - c i a - t a - -a - a - ‘iiL’ a- ’U i a - a - c a- 4 8K o r d a -a-r a-a-h ia - a - fo r a l l a - o r LO SS A t t u c a - n a - a - n n t a - a - a - t a - d i v r a - ja - -n cra -a - t a-a- ‘ a- ’ a - ’ a- a- a- w r c tt c n is n o t c l e a r bps l i n e s e e 2 4 K a-re _ ‘a- a-a- a-a - I i na-a- La - a - a - at it a- I a- “ la -a - a - a - aa- - i c ta - a - c TPL wa -s j u t e or u n r e l i a b l e Sa - a - a - a -r a - aa- - 1 La-a- i ’ DP a -j a-a- 11 t i e s a-a-ca-a- a-irrcu La-a l a-a- ov a - a - a - a - ca - a -ca-a - c a - n d ca -ca- a - a - a - i n c a - a - a - ’r v i c c aa-j -’ a-a-a- a--a - us a - a - a - a - w a - a - i da - a-s ‘a-a - h a -’ a -r o t a - c o l a- of dP S S -a- nd t a -i c Sia - j l a- I a - cu a b l e as a n a l t - a - n a t i v e to pr o c o s sa - a-a-a- - a - a -a- a a-- ca - a-a-r u ct cr s tl a - a - a -t a - a - u i a - a - p t c ra- a- a-a-a-i a- a - a- a - a-a -I v a -a- ’ its I r a - a - a a- a- h a - v a - c a - a - w er e i n t a - ’r r u a - a - t f o r a-m o t h e r a-a- a-i ra-a- a- a- er ’ a- a - a - J a - a - t O L W ore pa Ss ca -a a - a - a a-a-ta- ‘a - a - a - ca - aa- - r a- a- a- a- t a- a - wa - ta - a - to a-a- a-a - a - a - a - ta - ia - a -a -U a -a -a a- a- a- ’ a-a- a- e a - t a -’ c r a- a-a-a- t h e w a- r e a - a-a- a-a-d r a -’ a -a - a - a r- c to t a -c a -a - a - t a - a - a - a - i-a- - s t a - a - i”a - d a - i d a-a-a- j a a - na - a - a-a- use r s wa-a-en mca -dca -a-a- i n t e r — a- r ‘ _ _ _ a‘ ‘ a - ’ -‘ — a- - a -’ --a _ — 66 — Y uch p l a n n i n g and some c od i nq of thc ta-ad teen done during 1975 hcwever ca -a-il—level software grea te r clar it y fro m la t er documentation and a better understanding of the SWITCH interface n ecessit a -a-ta- ed some r e v i s i o n In addition it wa-a-s realised that the number of protocols that had to be supported and the desire to be able to add monitoring tools and possibly requ ired that a more rigid and well—defined programmed processes d e f i n ed Imp lenmenting interface between and the EPSS c a ll-con tr ol sof tware be this new interface a -a -a ll—control software justified da t agra m access involved reorgan i sation of the The delay that this caused was well by the ease with which the standard RL 360 driver from the 1 L AR PANET system was interfaced to EPSS This unconventiona l use of EPSS is discussed in the nex t section The standard interface is also being used for the File Transfer and Virtua l Call Protocols which are p la nned for implementation and for the V irtual Packet Terminal Protocol a- i e the protoco l for character terminals dialling up the PSE which wa-s coded and being debugged at the end of the year All a-a-a-c’ EPSS software runs under SWITCH and a - ’a - a - ac- ah- a - r a -e multi—network 6 4 environment could b e run i n a mul t i- The Packet Hasp Experiment Tha-” T’acket Hasp Protocol was developed jointly by the Rutherford -a-nd Daresbury Laboratories as the simplest m o d i f i c a t i o n that co u l d he made to connec t a standard multileav ing worksta tion to t h e Hasp p r o g r a m on the 360 over EPSS or using EPSS protocols over t eased a ll line The workstation uses a sing le EPSS call onto which t h e dat a for the card re ad ers line printers and interactive ta-n a- ’ta-in als is mul tiplexed in the standard way So far use of t h i s a- a-a-rotocol between UCL and RL has proved its viability and u s e f u l n e s s During the first quarter of 1977 we will also monitor the per formance of EPSS the Packe t Hasp syste m i s particul a rly sui t able for th i s b ecause i t keeps up a one secon d excha ng e of mes s ag es even in the absence of user traffic ‘ia -a-a-a- Pa ckct Hasp system became oper a tional d u r i n g the fourth - a - a - - a - rta - e r of 1976 We now run test sessions for an hour or so per ira - which I J C L s t a f f do rou t ine work v ia t h e U C L T IP and the r Ia -a - a - a - ARPANET Ep SS gateway At present the system requires a ‘a - a- ’ ’ _ _ _ a--- — ‘- ‘ a -— ” _ a t a - a- ’ _ a - i a- n’ ‘ a- ’ a-i ds is a- a- a-a- a-a-nra- I a - j’ i’t him t Ia -” r t not s u ch I a - n a -s a -K ‘a- ‘a - a- o t h e r h o a - a - Is a- a- a- - -a ‘-a- a- a- l a - a- ’ a- ’ a- a-a-a- a - a ’- t a - la-a-a-ta- i a- a- ‘a- a-a -ca -‘a- a-a- -m t a- a- ’ - a- ‘- t a - a- a- a- a- a-a- a - a - a - a a- a- a-a - - í a- c a- n a - a- - a- ‘ ‘ a- a - a - ’ a a - a - a - t -a a - a- - a - a - a - a - a - -a -a- a- a-r a- c- a- a -a - la -P a-a- a-itt a-na- - a -a- K ‘ a- a- a- a- a- ’ i ’ ’ - a -a - ’a -’ a- ’ ia - a -” ’t a -a- I a- - a -r ca - ‘a - ’ a- a- i S at 2 - It ia-a- a- r a-a-a- 1 f i na -it ’ ca - a - r n a- a-- a - a - t a-i a - l a - a - i a - a - a ta- P a - t a - a - k ’ t of La-’ a -p- a - a - a - a -ca - a-o ‘ia- ia- ‘- a- ste ad L a - a - a- a -a- a -a- a - a- sa- oa-a- c a - i a - i a’- a - a - ç a - r a - a- a - a - a - i r e a- ra- t 2 a - a -a - U U L s i a-a-t a - a - a - a- a - a- “a- la - a - a - a-a - c i a - a- a-a-’ a- S a--a- a - i l l a- a- - a-a-S i a - ta- a - e r a-Ta- fl a - - a - a - a - - ’l c- a - a - - a- ’ a-’a - a - t a - t i a - a - s t a -a ‘a--” a- n ’ a-n a-ic a- a-a- a- a - a - a - a - a -a - ca rt ’ ’ - a - a- il a - ’a- a- ‘ a - a - a -a- a- ‘ a- La-i’t fl ’ a- a - S S a - 1 a - ’a a - a - - a - a - it ’ 0 t’ I a - a - a - I’ c a-a -’ a-a-a- - a - a - a- a - a- a-a- fo a- ’ ‘ ‘ a- a - a a- ‘ ta - n na - a - a - c r t a- r a- I f a- ‘ a - i n q a- a - I a a-a- a- ’ ax ’ a- ’ a- a- -a- a- a- ’ a - a - c a - ’ a - a - s a - k ’ aa- - ’ a - a - n a - -i a - a - a ’ - a - P a - a - a - a - a - a - a - n a - a -a -a- ‘ a - a- a -a -c a - a - i f I a - c r 1 a-a-a-a- a-ca - a- a - a -’ a’ a- - ’ a - a -c a-a - i 7a- a - -a - ha - a - dule l a- a -a - a--V a-i r- ’a - a - a - a a- a-a- ta - a - - a - t ‘‘ a a- - L a - a - a ’- - “ a-a - a-’ “ a - a -’ - -a a - -I ‘ Ia- i ca-a- Li a-_ a- a- a- a- a - l a - a - a - aa--’ t a- ’ 1 a- a- ’ ‘a- a - a - c t I a - a - a a- - a - a- a-a-U a-a-a- 1 a- 1a - a - i ’ - a-ia-put a- a-a - roa IV - - a- a- a- a-Ia- ta-a -a-a- - ac a - a - a - a - a - aa -- ’ S a - a - rca - it a- ‘ a a-a- ‘a - a- ’ -a- ’ - a- a - - t a - a - fea - a - - -a-l - a- a - a- -a-a-a-a-a - c a- a - i a - a - ‘ a-a- a- a - t a - 1 a-a-a-la- a- a - a -a - ‘a- a- a- - i- - a -h a - a -a- ’ a-a- a - i ’‘a - a- a’ - w a- a- -- n a - a - r i - a - a - a - a -a a- a- a-a- a- a- ’ a - a - a - ’ a- a- i a- a- ma- a- I a- a ---’ l a-’ a-it a - _ a - _ a- a- _ i a - a - a - a- I t a-’- a- I a-c a - a - i l - a - a - a - ta IIq a- a- a- a - ’ a- a - a - a - a -’ l a - a - i ’ a- ’ n i r a - a - c i f l t t a - i l _ U a - a- ‘ a- a - a - -a- a - a -a - -a- a- a - a - a- a- a-a-’ ’ a- ’ L’ a - Ta - h aia - i a-a- a - a - i a - c W a - a a -- a - I C na - i a - a-ia- a- a- ‘ Ii 9 Ia- i a-- oj il a-l a- - a- l a - a - a- a v a - a - i ta - a - a -a - a - - or a-Va- -’ a - a - ’ a-’’ a - a- ’ a- a - a - a - a t a - a ’- a -a - a - ’ ‘ a- a- -a- wa - a -v ’ a- a- t ma- ia -c a- a- a - _ n ’ ‘ aa -a - a - a - - t u i a - a - a- a - aa- a-a -a a- a- a- a - _ a - a- a -’ ‘a - a - - a- ‘ a - a - ’ i V a - j l - 1 a - a - i I a - V o t a- aa- a- ia - a - ’ i ’ aJ a - a - la- a- F - a - a - a - a- - a - a - a- a- - ‘ - ’ a- a- a- ‘ a- I a - a - a - a - ia- ’a - to a-a- l ICC a- a - a - a -a- -ut b a - i f - a -i’ ‘1 t ’’ I a- a-a-a-a- I c a l l but a- a- a- _ a- in f a- ’a-a- ’ I ’ ’ a - ’ r a- a - a - a - a- a-ta- n L i ’ a - a - a - ” aa-i- the a - a - a- a- -n j a - a - a - f fa ’a- - ’ a - a - a - - i i ts a- n a - n a - i- a - a - - a - a - a a- -a nt - ‘a- ove r h e a - a - c l year a - a - _ _ ia - ri _ _ a-- a - l a - ‘a- a- a- a- ‘ -a - i l l a- a- traffic a-” ’ a- ’ t a - a - a - a - c ai ai a - a - - a- a- a-a- -p L y l a - ’ a - I ’ a-’a- ’t ’ sp a - a - a - a’- a a - e a- r a - m f f i c er a -c a - a - a - be a- -ra - r h a - ’a - a - a-i ca - a - li c a - t a - a - a - s u i t’ in SQl over tn c a - m a - i La -a -ca - a - a - a - S C a- a- t o-c - -a- ’a-y ‘ a- -a - i a - t a - ta - a -a- a-i ra- ‘ ‘a - I - _ _ Ia-i ‘ a- a- ta-a-a-f a - a -l a- -- ‘ ‘ I a - a - a - _ aa-i a-t r -a a- ‘ a-i a- a -’ a - a - t a - i fla- nw a - - a - - a - a - n’ a- a-a- a- a-c ia - a - of a-i aa- a - f ia- e ’a -u ” comm a- cout a-a-a-e a - ia - a - ring t h e a-a - a - y n c h r o n o u s T e r m i n a l s a n d FU r T r a n s f o r - ’a - l ’ a - a -- a-a -a t -a t a- ’ a-la- I a - a -’ m a - a- Va- - a - i p rot a - a -c o l P0 a- a-a - a - a i- l na - il ‘‘ l a - a - a - V a - - l a - l a - I a- a- a- t I i ’ s ¶ -a-c a - a - C I’ a- a-a- ‘ -a I a- a- a-’ aa- - i t a - - a- a - a - a - - i’a- a - a - ’ t a - t _ _ _ _ _ _ _ _ _ _ ta- t a - 1 it a-a- ‘ a-a-il I - a - i a - t a’ - aw- a - a -’ a- a- a- 1 a - i a -ia- m a - a - i a - a - - a a -- I nt - t’a- a-a - a- a - il a - a - li t a - a - i a - l a - a - a- a- a- a - a- na-ia - a - a- -a -d — I r a-i a- 19 7 a - w hic h is lt C’ t’nlii flq a - n a - I o ut a- l o i r a - a c a l l s f ’ ra - - ta - a - a- la - a -’ a - a - Wa - a - I a - Cu ‘i t t o a - i a - a -’ a - a a - a - a - ’a - l i k a - ’ i y to a- ‘-f a - I ’ a - ’ t a - a- a- l l ta--il a-’ I wa-I I ‘a-a- a-- a - ’ a- a - m a - a- a- a - a -‘ a-a-a- i a-’ ’ bli ’ a - a - w a- a- la - a - a - _ _ _ t a - Ia - ia - a - a - i t ‘ ida - a - a - t a-I i a- ‘a -a a - ta - a- La - ta - a - a - aa- -U a - a - a - t a - a -a p- a - ia - a - ‘ a- a-’ a - a - n a - a - a ’- ’ ca- -a - L i ” a- a - i i - a- - I a- a-a-a-a-a-a-i a U a- a- ‘a - ’ ’ a-” ira - a - a - a - a - aa -- a - a - -n t a - a - a - f a - aa -- a - a-sp ’c ts a- a- c n ’ a-a -n F u t a- i ” ’ a - a -ua - I a - a - r t a-i a- I i a - a - e O ba-a- f a -i t a - a - a - a - l a a-- — a - a - ’ ’ a -fl t a -- a - I a -’ a- i ia- -a - a - i’ ’ a- a - a - n a - a - I aa - ’a - a - i a - u i a - c a -’ a-a - I n a- a- a - a - a- a-a-’’ a-ca-a- a - a - a - a - a - -a - a - a -a-na - ’ a- a-a- - a - a- H a-c a- cia - a - i ‘ ‘ a - t a - a - i s m a -a - a- ’ ’ ’ a- a- a - h a-a-a- a - a - a -a- ia- aI - a - a - I ’a - ia - a - a - a - ia - a - ’ c a- ’ a- ’ a - a - v n a- - ‘- ‘ ‘a- ‘a - c a- a - t a - a a- - tm a-c l a- a-J fI a- ’ a-i a- ’ a- a - a - a - a- ‘a - n a a- — cx ra iia-c f u t u re p t a - ’lt a- ’a-a - ’ a - l f r ta - a - hi i’f a - a - ’r ca - t’ c ia -a -r y e a- ia - a-a- a- ‘a-s 1 11 a- - a-a- - terminals and that we may have problems with a-a-lapping other hosts adapta t ions back t o a v i r t ual t ermi na l However we anticipate betw een EPSS character in m a k i n g i n i ti al co nn e c t i on s no problems terminals t e rm i n a l s and EPSS h o s t s and h o ai a - an d ARPANET a- 1Ua-ANET We a l s o intend to support any v i r t u a l terminal protocols which gain wide acce p tance wi t h i n EPS S or which seem to have merit from an experimental view The EURONET v irtua l termina l protocols f a l l i n t o t h i s l a t t e r c a t eg ory poin of Iliggin son 1976D The SWITCH—FTP primitive a - of Section 3 3 are a subset of the sta ndard EPSS FTP primitives hence we antici pate no theoretical problems in b u i l d i n g a n i n t e r f a ci ng modul e b e t w e e n t he st a n d a r d process interface of the call—control software and t h e SWITCH- FT P Planning of this module is currently taking place and we expect it to be operational durin g the second quarter of 1977 For several reasons we have ordered a sec ond line at 2 4 1 bps to EPSS This second line will allow us to test cur gateway software completely It will allow also experimentation wi th the software neede d through concatenated networks whi ch was discussed in Chapter 4 We will experiment a l s o wit s our f a c s i m i l e t e r m i n a l a t t a c h e d d i r e c t ly t o EPSS v ia t h i s seco nd line see Chapter 10 When the h i g h level protocols have been ina -plemented on several other 131 hosts End we have ma pped the terminal and FTP protocols via SWITCH on to TELNET and the ARPANET FTP t he number of UK hosts accessible to A RPANET will increase considerably It is qui te possible at some time that the P0 and we may agree to res trict host access to computers which are attached to EPSS and not con t inue d i r e c t l i n k s l i k e the p resent Cu lha m RSRE and ULCC ones However before this can be con templated the a v a i l a b i l it y of EPSS must increase very substan tially 18 hours day - from 9am to 3am Even without probabl y t o a t leas t this h ica -h availability we expect to have a considerable oortion of the 131 traffic arrivin ca - via EPSS durin ca- the second half of 1977 We cannot yet g a-a-age the throughput of the U CL PDP9s running the multi-machine system under SWITCH so we cannot yet predic t on which UCL machine EPSS gateway software will usually be run a - - a -’ —a- - a- —- _ - -- ‘ — a--’ - - —— a - - a - -_— a- -_ _‘ - — t a- a- r a- a- - -a - a- a-a-t a - t a - a - a- la - a - a - i U - ul a- ’ 1 a- 7 ‘a-a-a- ’ a- ‘ a-a- a- a- a-a- a- a- a a- ‘a - ‘ a- ’ - ‘a - a- a- ‘ a- t a- a - a - a a - a -a- a - a-i i a - a - i a - a - a -t a -a -a- a-a- a- a- a - a-- a- a- a- ’ a- L a - I ‘ a- ta - _ i- a- I a- a - a - a - it a - ’ a - a -’ j ’ I a- a- a- I a- ‘ ta - I a- a - a - a -’ a- a- -a a- a - a - - a- ‘t a-a-a- i a- aa- a - a - a- a- ia- a-”a- a-a- a- a- i a- ’ - a- a- a-a -I a- a- a- a- -a- m a- a- ’ a- a-la- a-a- a - a - a - a - a a- - a - ’ ‘ ‘a- ’ a- a - a - a- a- a- “ a- ‘ U I’ a- r’ ‘ a- a-- a- ‘- a - ’ a- a-a-- t a- ’ a_- a - a - a - a - a -I a - a - - a- ia - a - a a- 1 - a - n - a- a-a- 11 a-a- ’ ’ a-i’ ’ ’ a- ‘ a- a- I a - - ‘a-a-a - f ‘a - a U a - a -’ I ’ a-” - - a -a- a- - a- a- a- ’ a-’ a- ’ ’ a- ’ t a- ’ iv a- — ‘‘ -a- a - Ci a- a- a - a - r - ’ a - a-_ ‘ ‘ - ‘ i a - a- ‘ a- a - a- ’ a- l a - u ‘ a- a - a ‘ a- a- ‘ a- ‘ ‘a- -a- a- a- a - a - a -I a - ‘ a- ’ I a- ‘- a-a-a- ’ ‘ a- I a- a -I a-a-a- a-a- ’— a- a- a-a- a- a- ‘ a- a- ‘ a-a- a - a - a- a- a- a- a- a-”-a-r-2 ’a- ’ I - ‘a- i ‘ l a - a - a-a -Ia -a- C a- ‘a- f a - ’ a - a - a -a - a - ’-i a - a a-- L a- “ a- i a-aa - a - ’_ ‘‘a-’a-a- a- a- - c a-a- a- ‘a - a- I ’ I a - a- a-a- a- 1 ” a- a- a- a -’ a- ‘ a a- a- r i a- _ a-l a- I -a a- a- a - I — — ‘a -’ a- ’ a- ‘a‘ a -a - S ‘a- ’ ’’ a- a- a- a-a-C ‘a-a - a - i 0 a-’ ¶ Ir a-- a-_ aa-” a-a-a- ‘ i _ a- l ‘I a- ’ a- a- a- ’o’’ a - a - a - a - t a - C a - a - a- l a - i a - ’ ’ a -a- a- ’ — a-’ a- a-a - a -a - a- ca - f a- a-a-’ a - a - i a- a - a - a -a--a- “ a - a- a - a - a - a - a-a- a - a - - t a- a- ’ ’ i a - a - a - a- a- a - i a- a- a- i s a-a- iL-a-’ a- iN I ’ a-i a-a- a - ’ ’ a- a- I ‘ a- — a- a- a-a-a -a-a- it a - a -’ta- ’ -’ a - ’ a - ’ a- a -a- 1 -7 7 a-a-a- a- ‘ ‘ a- um’ a - n a - i T T a - a- a- a- a- a -m o a - a - a - s a - a- a a-- ” a- a - a - a - a - n 1 Ca- 1 ’ f la- Ca -a- IC ‘a- a- ‘ a - a -- a - a- ’ ’ ’a - i’a-l a- a - a - ’i - - a- a - a - a- a - a - ’j ’ a- a - L a - 1 a - a-c t - ‘ a- a - a -_ a - a- a-v a- a- a- a- a- a- r a- c a - a - a - a - a - a - I a - a - a- ‘ la - ’a-a- a-’ l a - a- a- I’a- t’a- a-a- a- ’’ -aa - a - a a-- a - a - ta - a - a S a- a- a- a-a-a- I a - a - a - i a - a - a - ha - I a - a - a- a-a- a - a - a - a - $ a- C-a-’ ta - a - a’ - a-’ ’ a-’ _________ l ‘ -c a- a - a - - t’ a-’ a- a-a-a- a- ga - a - a -’ a - I a - ’ - a - ’ ‘0 - ’ ’ a- a - a -’I a - a-a- a- a- a- a- a- I a- a- ‘a - a-a- a - a- ’ L i a - a- ’a -’ I a- - a - a-- a- a - a - Ca-a- a-a- a- a- a- a- aa - a - i a - a - a - a - - a - a -- ’ ’ a - i a - a - -a a- a- ’ a-a-a- a- a- ’ a a- a- a- a - a - Ca-a- a - a - a - ta - a - -a- a - a - a - a - - n r a- - a - a - m a- t a - t a - i a- a- - a- I _ a - a - a - l a - i a -a -a-i t a - a - a - ‘ - e a - a - J ’ a- ‘ ‘ ‘ ‘ a- a- a- ‘a - a - I a- a -’ ’ a - a - a - a - a- ’ ’a- ’ i l ’ ha- ‘a- a- a- a- t ’ a- a- ‘a- a - a - a - i a _ - a - a - a- a- ‘‘ct a- a- a- a- -a a- ta - a - a - a -a - a - a - a-a- ‘a - a - a - a- a-’ a- a-a-a-a -a- a- a- “a- na - ’ a ‘ ‘ a- -a -a - a - a - a -a- ta - a - a - a - n a - ’ ‘ - a -’ ‘ ‘ a- a- a - a - a - a - a a- a -- ’ a -a - a - a - a - aa - -a - a - a - a - a - a - ia - a ’- a- a- a- “a - 1 a- a - a - N a - il t a- a- ‘ I a - a - a - a a-- a -1 I - a - a - ’’ I -I a - i’ a - t a - a - a- -a ’ a- a- ’ a- a - a - ’ a- a- a- a- a- a-L aa- ‘a - ‘ ‘ a- a - h I ’ a - a - a -“ I ’ ‘ a - a -a -a- a - i a - a - a - a - a - a - a - a- a- a- a- a- a- I ‘ a- a- ia-a-a- a- ’ ’ a - t a- a- P a - ’ ‘ ‘k a- 1 a - ’ i a - a - a - a - a- a - ’ a - a - -C ‘II - a- a- a- a- ‘ a - a- ‘a-a- a- a- a- ‘ ‘a - a - a - a- a- a -a - ’ -pa - i a - ’ i t a- aa- a- ‘ ca - a - a - a - a- a - a - ’ a-’ - a- a - ’ a- a- a - a - a - C a -ri a - a -’’ a - a - a- a- a - ‘a- a- a- aa- - a - a - a - a - a - a a-- a- c aa- a- a- a- a- a -’ ‘ a- — aa - a - i t a - l a - a - - ’ ’ t a - a - - a - ma - a- a -a - a - ‘ a- a - va - a - ’‘a - a - a - a - a - a - a - a - a-a- - a - a- ’t a-” a - i a - a - a - ‘ a-a- a - a - a - a - a - a-a-a - a-a- - a- a- a-t - a - I ’ a-a- 1 1 L a a--a- I ‘ a- a-a-a- a - i a - i a - Ia -— ia - a-a-a- t i r a a - a - a -a - a - i ’ a- a-- a - i a - ta - a - a- a- a- a- a - ia - ’ a - a - 1 a - a - a - c a - a - a - a - a - aa-a--h a - a - a a- a- a - a - a - a - i a - a - a - L a - a - a - a - a - aa -- t a L a - a -t a - a- a-a- — a- ’a - a- a - a - ‘ a- la - a - ‘ a - c a - a a-a-- ta - a - ’ - a- a - a - a- a- a- a- a - - a- a- a-a-a-a- t w ’ rca -a-a--a-Va-it a- a- a-a- i i - a -aa - a - a - a - a - a - m a - a ”- a - C La - a- a-a- C - a-i -a I a- - a - a - a - a - o a - a a- -’ ’ a- 1 ’ a-a- a- IL t a - a - a-a- - a - a- a - ’ a-’ a- ’ a- a - a - ‘ a- a- - ‘‘ a- a- a-a- a - a - a - i a- - a - a - a - L a - a - a- i ta - a - a- a - a - a - a - a - a-l a - a- - a - a - h a - a - a- ‘ a - a- ’ a- - a- a-_ a- a- - - a-a-a- i a - a-a- a- a- ‘ a-a - a- a- ‘ a- a - a - L a- a - Va - a- a- a- o r a- a - a - a - ’ a- a- a- a- a- t a - a - a - a - a - 5 1a- a - a- ’a - a - a - a - a - ’ a-a- a - a- a - a - a a - - a - a - a - a- La - a -ta - a - i a - a - a -a a -- a - t a - a -a - ’a - ’ j a - a’t - ‘ i a - a- a -’ a-a- t ‘ a- a- a - i a - a - a a-- aa ’- a- a- _ a - - t a-ia -a- a - a - - a- a-a-la- i L H n a - ’a - a - a - a - a- - E U a - ’ a- ’ a - a - a - ’ a - ta a-a- ca- ai - a - a - a - c a - aa-a-- a- a- a - a - a a- -a- a-’ a- a - w — - -a - - - - ’ — I a - a- t a-a- a - a -a- a -- a-a-’ a- a - a a- - a - a - a - ia - a - a - a - a ’ i a - aa-a- a-a-’ a-a- a- a- ‘ ‘ a - a - - ‘a-i ia - a - a - - a - a-i a- ‘ -- — ‘ - a- ‘I a--a- ’ ‘ a- ‘ a- a-aa- a-a-a- a- a- a- a-’ ’ a -a-a- a - a -a- a- a - a - a - ‘ a - o s sa - ’a-a- c a - a - a-a ’- a- a- -a- _ — — _ _‘ a- ’ — -— a- a- - a- p•— ’ a- a wa- a- a-_ a - ___ ‘-‘a - - It is indeed possible tra - a - t other sat a - t’a - l l i t e s t a - a - a-a-a-c other satellites ‘i n c l u d a - r s j the European Orbiting Test S a t e l l i t e ia - y he u s e d ira - t h e However the present project is a - -o a -a -cerned onl y w i t h use a-a-f INTELSAT IV a d with PANET s an a c c e s s n e t w o r k the future The detailed objectives of the project are discussed Linkab it 1976 elsewhere Because the project involves many p a r t i e s A RPA - BBN COMSAT DCA LINKA BIT the UK P ost O f f ice and UCLA it is to discuss these here c leaL ’ly not appropriate report it is enough to describe this 7 2 arid the general progress during For the purpose of the actual confi guration Section 7 3 1976 The p r i n c i p a l UCL activities in the Packet S a t e l l i t e been to develop tools for traffic the configuration of Section 7 2 Project have generat i ra- and data acquisition for We have concentr ated on hi g h l evel generation measurement outs ide SATNET i t s e l f traffic Section a -vents data w i l l also be provided as the traffic though measu re— — traverses SATNET While we have been able to develop the tools in the absence of computers arid other sites no real UCL measurements will be Gateway made before A p r i l 1977 Sections 7 4 7 6 t h e ma in su b j e c t of to The genera l description of these tools is t h i s ch a p ter and they are discussed 1976 Our progress d u r i n g in is s ummaris ed in Section 7 7 how the tools will be uti lisea -I is t n e s u b j e c t of Section 7 8 B ecause the UCL measurement tools are d i r e c t e d at h i g h l e v e l a-a- ea-a-surera -ients they are almost independent of the network being measured C l e a r l y the packet fo rmats and the exact timestamp mechanisms are by t he data ne twork being inves t iga ted a- ov er n e d par a-irna -et ers w h i ch must be set are a l s o d ependent A l s o the e xa ct on the data networ k aFor examp le there is no point in setting priority a c i 1 a - t i e s I a - - a - i p a - t i n a t i o n or m u l t i p a c k e t t r a f f i c in networks w h i c h c a n n o t i’ a- ‘ i a - a--a-’ The a p p l i c a t i o n of these tools to SMa-NET w i l l involve a- ’ ’ aa- a- i u r a - k a -y SATII4ET switches of additional i a- a-a-a- r a - r ‘‘“i a - i - ‘I -a ka - j a- - ‘s ‘ ‘ a- ‘ the packet inforrna t ia - a - n on their ira -formation switches in the n e t - a-a - a - a - i rk would have itself ra- i a - a - a - r a - tion • a- - net works s i m i l ar other -a 1-ri a -v a -ty b i a - a -t ’ — ARi ’ a- i t is as su med her e t h a t t h e re ad er is a -NE a - I ’ tec hno l o g y and t he f unct i o n i n g of — -- ‘ -‘a- _ a- a- a _ a- a- —a---- a - a - toe Sat e li te IMP s B u t t e r f i e l d 19 7 4 In this c o n f i g u r a - a - t ion is own in Fig 7 1 no r e a l f u n c t i o n — t h e S a t e l l i t e IMPs SIM P s a r e s t i l l connected to t h e IMP s or a - l i P s of way IMP— IMP It t he c o a- f g - a- a -a-a - r t ’ t a - a - c e a - a - r t a - a - 5 t 3 t 1 a - ’a-’a- S ARPANET ta - - a-c in a-a- a-a-rn - a - a - i ta -a - at t h e y ia-ave s p e c if i c wa y s of c o m m u n i c a t i o n v i a t h e sinq l e s a t e l l i t e c h a n n e l i a-a- teway Computer h a s two addresses in the A R P A N E T a-a-a-st 234 the fourth local host on t h e CCL T I P ha n d l e r of t h i r d t h e TIP as the and h o s t 124 h ost on the Goonhilly SIMP which a- iMP for routing purposes Th e a - ’-a- t h N a - a - r w J il’ a n d t h r ou a -a-j Ii norma l traffic if Fia-J S IM Ia- 7 1 - T h e 5 the t IP a t l a - a ’- t h e c i r c u i ts two n ot a- a- a- a - t a - a - i l l to t r a f f i c b ut p e r m i t s t hi s In La - a - c first Very Distant if dS and ETAM has US n o r a - a - a - al were an it — Lo n do n G o o n hi l ly avoids the disturbanc e of a - a - a - a - a - a - a - a - a - i aa-a-a-a- a - a -aa -- n t experiments by This na - a - a - na -a - il c o u n t i n g t h e t e r r i n ai be tw c ea -a - O o o n h i l l y l a-ta-en ia-iade to appear a-a-rt ificial y long so ARPAN E T t r a f f i c passes The CCL routin g tables it ia- ARPANE’1 ’ tra -’a -a -t s p at h Gateway 7’a - Pl l ha CCL ia -a- is o n l y b e t w e e n each o t her a - a- r e s e n t PSP ‘ia - a - a - c a-- a - ’ ch a - i a -’a - n e l t h ’a- ’ a - u a - h N a -a- a -a - a - wa - a ’- a - lines are — to ha-’ used f o r a-re down s h o w ” b e t w e e n CC L kbps one terminates ‘ ‘i’A t he t h e G o o n h i l ly a- a -n d t h e P U P i l t h e 4 6 iKbps one or a- a-fl The second line ha-a- S b e e n a - -a - it i n a - a - a - a -Ia- ’ t e n a - p o r a r i l y it is to permit diagnostics and control of t a - a - c a- a- c a - e nh i i l y S Ia - i P t o c o n t i n u e v i a N o r w a y f r o m t h e N e t w a- rk C o n t r o l Cent re even if there iS difficulty with the satellite channel w i t h i n a-a-a- a-a -a- a-c contention stratea- or with some h ar d w a r e m o d i f i c a t i o n s Th e conf igur a - a- it io n wi 11 the second a- ’ T h e re w i l l c i i a -a-a - n a - a - c quarter of 1977 a- a - a - a - v a -aa -- a - i in new n a - la - a - a - a - be two additiona L S1MP5 — a- a- a - a - a-a- I i a - u r a t i o n h a s a smaller dish INTLLSA T IV a- a- FAa--I is a - l a -t a- i t Na - i t t a - a - c i a - ow on a j c i o r an d a - il it a-a- Ne twork in if necessary l ab o r a t o r i e s IU M -’ a- a -n ad d i t i o n a l ‘i ’l a - e a E3BN a- it a- ii ra -u via t he access line from in Cambridge A s i m i l a r conf igur at i a- a- a -aa-c Now t a - a - a - a - will be a rea l a - a - a -’ F A’ a- ’ it t i - a - a - cd a - A a - i a - N I ’’a-’ a - u r a - a - a - a’a-- a - a - ia - l y a - r i d ‘1 a- 1 la - a - ‘a-a- w i l l a g a i n la - a - c I’Dl’ 1 1 G a t e w a y c o mp u t a- a-r i m a - a - n a - i a - a - I s a i - a - n a- ’ a-r ’a - a - L n l ’ a - a- i a - ia - o re T in F i g a pp r a - a - v a l to access b u t s t i l l w i th ua- -ing shifted to terminate a -a -iso a - X ia -i t a- ia - a -h ot t h e COM SAT SIMP t h e COMSAT one nia -a- y be c o n t r o ll o d t e l ep h o n e network ‘a - a - a - i sa -i ch u s c t t s wi H — sa t a - e l l i t e t o control the Tanum ‘ a-a-iit chc d shown ta -a-a-a-se is another sL -a-nda rd INTLLa-SA ’i’ e a r t h station t h u ‘l a-a-c I a - r s t of the is at ta -a -c N ord ic earth station i n Tanum and t h e COMSAT e a r t h s t a t i o n at a - a - c a - f l a - I I a - a -a - a - o r ta nt as p e c t s dur i a -a -a - t i i r o a - a-qla- ‘l ’ha -’ ‘a - t a ’-t a- ’a- ’I ’ L 1I C ’ t a - a - A a - l’ANia-a - ’Ia - PDI’ u s a-a-1 1a- a- 1 w i l l a - it for ia-a- ”a-’a- a- iC a’ -n be m ttached also to 363 f o r nle a - isuroii a - a - -niL d at a -i a c q u i si t i o n and a n a l y s i s - a - a- - ”-- - ‘ - _ — ‘a- ’ - -’ - -- ’-’-’- - ’ V a- -- ’ a- ’ I 0 - - - ’ -“ w I— ’ -- a - --a- -_ a - z Z 0 OL L 0 0• _J L1 4J LI _ 1 w a-a - a- 0 a- - a-a- - I I O D I t I I— LI z a- Z -J I ta- a-- h’ W a-a -’ — LIa- C a- — ‘- 4 a-— ‘a- 11 r LI 0 0 N Z Z a-a- 30 3- z z 8’ 0 3- -j 0 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - “ - --a- ’- w —- - ‘- a - — --_ — -- ‘ - - a- “ — -a-a--’a--- a-- - a- - — The additional 9 6K bps access line to the SIMP s are being retained purposes they w i l l be removed as soon only for test and diagnostic as BB N has perfected the techniques which w i l l a l l o w t e s ts an-d diagnostics to run throug h t h e P a c k e t S a t e l l i t e c h a n n e l The new configuration will radically alter situat ion the experimental SATNET being a four earth station network wil l allow much more r e a l i s t i c tests of contention strategies with Because i d e n t i c a l gat Ew ays w i l l be attached to a l l w i l l be possible to have s y n t h e t i c t r a f f i c it Finally da ta a c q u i r e d in the Ga t e w a y s the c h a n n e l t he e a r t h s t a t i o n s generated it w i l l and measurement be poss ib le to change the protocols between SIMPs and between SIMP and Gateway without any need to remain consistent with ARPANET procedures At present no f u r t her chan g e in hardware co n f i g u ra ti o n is pl a n ned d u r i n g t h e l ifetime of the experiment 7 3 General Progess During 1 97 6 Th i s is the annua l repor t of the U CL group and not the other PSP participants For this reason it is not a p p r o p r i a t e to give lenca- thy discussion here of the progress of the other groups During the year there were p e r s i s t e n t problems with the s a t e l l i t e channel itself particularly in the contention modes These were traced to stability problems in the SPADE experiment under the novel require- me nts impo sed by the exper im e n t Se v e r a l m od i f ic a ti on s were made both to the earth station modems and the Spade SIMP - interface SSI these have alleviated the situation and made the c h a n n e l u s e a b l e for experiments a- In order to keep c l o s e control of the channel a n d t rack down the reasons for problems e x t e n s i v e d i a g n o s t i c s were buil t into the Extensive diagnostic SIMP information passed back to BBN via ARPANET to b e a nal ysed t h e r e statistics are then passed to all the interested basi s t F a- is now The c h a n n e l parties on a daily The whole d a t a a c q u i s i t i o n analysis and d i s t r i b u t i o n In addition new SSI units da ta is done automatically developed by COMSAT and will be fitted at a - ill the Tanum and COMSAT S IM PS are commissioned of are being e a r t h stations when The whole area of diagno s t i c s and maintenance have rece i ved grea t at t e nt ion from BBN and has led to the provision of the extra access lines of Fig 7 1 a- There have been extensive measurements by the UCLA group of some a- of the simpler strategies for using the satellite channel _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ These -a- ’ a- -’ - --- ‘ a - a - a - Pa - V a - a - a-- - _ _ _ - — a-n described in reports i ’roa-a- a-a- a- a- a- a- a-’ E x t e ns -a- -e ck’ s group ra -ia -’inro new contention strategies for usia -a-y th€ satellite channel have been carried o u t a-’a - t BbN L L N K A B I T a n d U C L A d e s a -y n and s i m u l a t i o n s of ‘I’hcse strategies a i m at usin g the chi a -i r a -n e l e fficiently an d - th a - a - d r ecovery f r o m er r o r s overloading and with different traffic n xes also a-llow for the first time the attachment of priority and They maximum permissable delay on ta -a-c dat a-i pa - a- c kets Both of these be required in the commercial or military environment Implemeri at n of these strateqiea- in t h e S I M P h a s been s tar t e d by BEN w i t h t h e a - a - ia - i a - a - of deployment during 1977 Details of the above a-ire contained in a-i notes the Packet Sa-itell ite Project W o r k i n g may be available ta-a- notes themselves s e r i e s of i n t e r n a l w o r k i a - a - -a a- Group notes interested parties from tla-e a u t h o rs of One gateway P U P i l h a s been i n s t a l l e d at t h e UCL IMP and also on the Goonhilly S F-a-P These the UCL as an ARPA N ET h o s t or da-iteway code has been provided by BBN so ta-a-at incoming packets with a particular Internet header as used in Chapter 5 can be timestamped and passed to a There are a-a-lso facilitie s f o r c c c r a - n aprogra -irra - b e i n g developed i t UCL a-nd forwarding packets a-a -ca -a -era -a -ted b t h e UCL nr o g ra m to either t h e IMP The host to ARPANE’r at 1113N be described will a-a-a- S 1MP a t t a c h e d a- a- a - a- a- functions in Section 7 4 a - md of these UCL progL-am s A simil ar P U P i l ca-i n h a s been at t a c h c c i be u sa-a -d f o r t e s t i n g t h e UCL a - a - a - a-a a - -c The DCL Measurement Tools 7 4 A - An O v e r v i e w high level traffic genera -a -ta -a--r m easurem ent tool i e one f ru i t o u t s i d e t h e n e t w o r k i t s e l f r e q u i r e s ta - a - a t t h e n e t w o r k u n d e r measurea - a - a - nt be seen and interpreted from a host level point of view h a s two im m e d i a t e c o n s e q u e n c e s ha-a- -a- r e a l i s t i c model for t r a f f i c F i r s t t h e experimental traffic must qe n er a - itea - i at host ieve l m e a s u re m e n t a- must be o r i e n t e d work perfa - rma rice level fi gur e s as network Na - a - a - th e r of a f a - a -r While it variables r c a - a - a - a - a - a - a - a - a to -c m ay This la-a-’ level second hos t towards a-i user measurement possible to e x t r a p o l a t e of net- from lower such a - iS t h r o u q h p u t some var r a - i bles such u s e r demand can o n l y be measured a-a-i t host leve l tha - ’ a - c requirements depend significantly on t h e f e a t u r e s of p art i cu li r network a - m d accordingly much of the following discussiora- is couched in network independent terms in any particular implementation -- ‘-‘ V a-a -- ‘a- - th e ne two rk interfaces need clear specification a t ool provides a useful means of assessing On its o n su ch the o ”era l l p e r f o r m a n c e of a network but would not tell us much abou t the importance of any single feature This type of information can be derived by f e e d i n g ‘ r e a l i s t i c’ multi—user traffic through lower lev el mea uremen t tools in an integrated measurement scheme Bearing these considerations in mind there are three important aspects of arm integrated high level tool viz acquisition the generation and of the tr affic to be measured the contro l and variation of traff ic generation in the experimenta l environment and the actual measuremen ts to be made and 7 6 These are discussed in Sections 7 5 Three main traffic patterns are recognised in this experiment each of t hese t r a f f i c pa t t e r n s may be var i ed so th a t a w i d e spec t rum of d a t a types is a c t u a l ly covered 1 The th ree main patterns are High data rate bulk transfers Low data rate stead ystream t r a n s f e r s 2 Interactive traffic 3 In the SATNET protocols to be im p le m ented i n 19 77 ca tegories 2 and 3 are trea t ed as s i m i l a r t r a f f i c type s for purposes of sa t e l l i t e channel reservation ho w ever for all traffic Priority and delay ca-a- ri be specified An example of a high data rate transfer might be that of a file a c r a -a- s s the network The message generator program would attempt to push messages as fast as possible into the network g e n e r a t o r m ay t ry to t r a n s f e r a prescribed n umber of The b y t e s t hrough t he net and measure the tota l elapsed time or may j u s t t r a n s f e r as much data as possible in a given amount of time In e ither case the sys tem will measure the maximum throug hput th at the user could expect Combining this traffic with other controller traffic on the net sec below will indicate how the performance deteriorates with interfering traffic The low da ta rate steady stream transfer covers such data types as voice traffic where the tota l amount of data may be small per unit of time compared to a fUe transfer but the messages enter _ _ _ _ _ _ _ _ _ _ -— - ‘a- ’-— -- - iL very j a-rec i a - a - a-a- inta-erva 1s Once v o i c e the network a- a-’ nta - a-i de livered n ot only m u s t be a - it a-a- coa - a - st a - in t r a te h u t a l s o a- ’ ith l a- a- t a-a-a -a-ay- a-a- s p e c i f i e d time period every ha lf if for a - a - a ’- a - a - a - a - L e s instance second it must be d e l i v e r e d message enters the net a-’i every half second at the other site ‘ bunch in g ’ of data is not permitted here unlike bu l k transfer Much real time data s uch a s vo ice f a c s im i le telemetr a - a - a - a - e me a sureme nt s or control data falls under this category ili consist of examining how consistently the messages are delivet€a-d to t h e d e s t i n at i o n and t h e n u mb e r of messages that arrive at the dc a-tination with in a g i v e n t i m e ot entering etf a - ’ct s if ot a-a-er traffic in ta-a-c n e t w i l l Interactive ficult tra - m f t Ic ia - a - t h e measure most d i fi H ere a -y st e n i the too t a - a - e Ia-’ a-e a -ts u r a-h ’ t o s i m u l a t e and a - cult the ta -a -c user sends short massa-ies of variable size at variable intervals a- ir a-h receives r e p l i e s con sistin of o n e or more l i n e s a short t i m e l a t e r T h e important aspect of p a - a - st dit t a- t h i s type of traffic a- 1a - y p i c a - a - l ly i s w h e t h e r t h e n a-ess ’a-a- a-e s can be d e l i v e r ed w i t h a - n some threshold time t r a f f i c g e n e rat e d The ba -ised on s t a t i s t i c s q a t h e r e d on u s a g e a- messages sent and t h e t y p m af ‘ i a - a - a - a - a - t a - j a - a - a - a - a - rem ui ent t h a- a - h i I i t y 7 i t- of t o h a n d le al l 7 ’j I a-a-a- p r o vid e d w i t h tla- e n e t w o r k I e r i n a - e n t s has an thr a - a - a- m a i n t a- a -a - y by a-i SIa - - T l e v el t r a f fi c qcn a- ’ r o o a - a - ion t h e p r o g r am t h a t sia -a-u lat e s users ’ ir a -te r — Each g a t e w a y ta - a - k a- a-a-- a- p a r t in t h e s e ex— identica l copy of s egm e n t s t h i s m o d u l e wh i ch c o n s i s t s of s e g m en t t a - a - c ha t -‘a- g e n e r a -it i o n seg ’ne a-t the control a -a- r e d u c t i o n s eg m e n t a- a-i a- it a- a- a-a- a -a - it ra - I S C q f li e fl a - -a - ’ q-mt ewa-iy t - ia- a-ce a- t S a- a- Oi flfl la -I na-Is f ta - a - a -a - a - a - a - a - fl t a-l ie a - t a - e l ’ Sa-a-’ a - ia‘a- nd a - a - - a - a - a - r a - i t u r s i m u I a - a - t e s L h a - t ’ U s e r ’ S i n t a - ra - ma -’tior i a - in ot U ’ ea - a - s i l u ’ a- a- The Env i r o nm e n t a - a - c t ia -n a-a-a-t i t y p e of a-a -a- is t h e rn a-a- or de s aT a - a - i s r e q u i re s greater f I e — D a - a - t - a - i c q u i s i t t a - Ia - is Ia- a d e s c r i p t i o n of a-nd these t r a f fi c a-he t r a f f i a -- g e ne r a t o r Ca- a-fl ta -a -c A R P A N E T t h e r a - a - p l i a - e s r e c e i ve d t T r a f f i c Gener a-i t ion and of by t h e p r o g r a m w i l l oc a-a- a-nt i’ a - a - l s ex t a - a - a -’aa-- t’ S aa - a- a- a-a-a-a- I ot t a- h e I a - a -a - s t p r o q r a - m a - - ’ ir a- t h e f o l l o w i n g ’ a - l i ia - ca -a -aa - a - a t h e word user w i 1 1 d e n o t e a-i s it u La - a - ted u s e r M essages aa-j cn ’ r a - m t ed i n d e p e n d e n t l y for a- a-m ch u s e r a n d t h e m e s s a g e s a r e t r sr a - s rc n ra - It ’ d a -m ft e r p issinc through a-a- simp le I l a - a control mechan ism ‘ a - a - a - a- - a - il ‘ ‘- —a- -a- I The data reduction segment will accept mess ages from any gatew ay taking part in the experiment Pertinent s t a t i s t i c s selected b y the ecperimenter are extracted from the received user messages and ga thered in histog rams are sent back 7 5 2 to a data At present intervals the histograms collection s i t e The Traffic Generator The generator is controlled by a User Traffic Table which contains the message generation details for each user being simulated Each entry in the table describes a separate user an d at present the table has 32 entries allowing up to 32 users to be simulated from each gateway at a time The user traffic table is loaded by the con tro l module The table can be loaded or modified only be tween exper iments when the message generator is not using the table Each entry in the table contains all the information required to s i m u l a t e a user although the number of entri es arid their actual value can be given For instance the experi menter can specify tha t for a particular user messages are to be transmitted eve ry 10 seconds 2 seconds The generator uses a random numbe r generator to select the actual value within the spread constraints The most importan t parameters that can be set are shown in the a- following table U s er Numb er A number from 1 to 32 Mean Message Size Mea n Size of gene rated messages Message Size Spread The Variation allowed in message sizes bytes Mean I n t e r Message De l ay Inter Message De l ay Spread The mean delay be tween sending each I ext Message N umber between messages Length of the next messag e to be sent 3 cheduled Departure Time The time at which the next message will be sent t raffic Type ‘f able 7 1 The variation allowed in the delay messages N ex t Message Length Nes tination message A sequence numbe r inserted in t he Interactive steady stream or file transfe r - - In The gateway these messages will be sent to The Important Parameters of the User Table All the entries marked with an loaded by the control program ‘ ‘ are the entries that mus t be a- _ _ _ _ _ a-- ’_ - — i use r ’ s mess age is transm itue d and at the b e g i nn i r ’ a- of the e xpe r im e n t the data g e n e r a t o r g oes throu gh a c a l c u l a ti o n Wheneve r a to determine at what time the next messag e for that user shoui be sen t and what leng th that message should be As soon as the Scheduled Departure Time for a message is equal to the current clock t ime the c’enerator attempts to transmit the message h oweve r this attempt may fail if the gena-rator determines that t h e ga teway is un able t o accep t the message the user entry is put a frozen state until 7 5 3 the gateway ira - is ready Expanding the Message Up to t h i s p o i n t the message h a s c o n s i s t e d of o n l y a few e n t r i e s in the User Traffic Table The message generator actually manipulates pointers to the table rather than moving potentiall y massive mess aqcs around The m e s s a g e is e x p a n d e d to i t s f u l l l e n g t h j u s t b e f o r e t r a n s m iss i o n Eaca- message 3lways contains the following information S ource G a t eway Nu mber User N u m b e r Message Sequence N u m b e r M essage Le ng th S cheduled D epar tu re Time Actual Departure Time T r a f f i c Type Destination The e n t r y m a r k e d departure is i n s e r ted a t t he i n s t a n t of t h e m e s s a g e Af ter these fields have been inserted the message is expan ded to its full length with d umm y characters Special consideration must be g iven to the steady stre am traffic type For these messages It is important a threshold tun e If that they are sent w i t hi n the d e l a y p r i o r to t r a n s m i s s i o n exceeds some threshold value the message mus t be dele t ed th ere would be no poin t in t r a n s m i t t i n g i t as i t is already out of date c ase the message is deleted before transmission In this and the generator proceeds to calculate the next transmit time for that user a - _ a - - a - For -— ‘ -a- a- a- a ‘ - st a- r a -a-a - c a - a - a- t ’ a- ’i f f de -a r t’ a - - a- I c the delay - to the next messa-ige ia -a-a - based on the scheduled t ime of the previous message ig t h e Scheduled D e p a r t u r e Time of t h e p r e v i o u s m e s s a g e as a- a- s t naa- t h e b a se time w h e t h e r the p r e v i o u s message was c a n c e l l e d or n o t a - a - a a-i- r a - s t a - a - at messages are generated at regular i n t e r v a l s r eg a r d l e s s of With other t r a f f i c types Lra-e a- cceptance rate of the network where the current clock tirreis used as the base time t h e i n p u t stream ca-i n ‘h i c c up ’ when the network is u n a b l e t o t ake a n y more traffic a - ’ a - lri • ia- consistent with actua l network use If t h e ne t w o r k ra - a - Iu s a-’s e n - a - a - ’ja - a - to accept a message because the destination h o s t is not r a p i d if or t h e IMP c a n n o t accept more t r a f f i c t h e u s e r g e t s any ca -a - z r ic s r e j e c t e d t ry However this simply causes the a g a i n n o t to d i s c a r d the messages he w o u ld a-a -a-er to wait and h a v e s e nt This is a l s o t r u e of f i l e t r a n s f e r s w h i c h a l w a y s t r y to t r a n s m i t as f a s t as the oa - a-tw a -r k 7 5 4 Data Acquisiton and Reduction will accept n t h i s segment messages are accepted from other gateways in ta - a - c experiment and the data is reduced before bei ng sent on to the The data reduction segment n ever denies entry t o CO 11C CL a - 1a -J h o s t t r a f f i c from the n e t w o r k It w i l l at a- an y t ime i ra-a n any sequence accer a- t messa oes from any aateway As u s e r m e s s a g e s pass through the sa-itel lite network to the destination la-ices gateway they a rc tim estarnped at up to ei a-a-ht different By inspecting these timestamps at t h e destination the time a-p er t by t h e message in various parts of the network can be determined These times can th e n be e n t ered in h is tograms To perm it ta-a-c yreatestflexibility a g e n e r a l h i s t o g r a m method is employed The experimenter can inform any gateway via its control module which users irom wh ich source gateways are to be histogrammed The experimenter can al-so select the tim estamps terested and which of the modes are desired in which he is in- In one the ‘ successive ’ mode the difference in a particular timestamp between successive mes s ages is us e d in the o t h e r the ’ a-a - hif fe r ence ’mode t h e d i f f e r e n c e bet a- en two ti me st amps in each me s sage is use d Each histogram can also be given a ‘l i f e t ime ’ A lifetime of one m i n u t e for instanc e wou ld mean that every minute t h e histogram -4 be sen t back to the collecting S i L o 3nd then re-initialised t o a - h a - s a soric a- ‘ f zeroes for subsequent statistics collection histo grams for a user can give a ‘mo t ion p i c t ure ’ v i e w of t he us e r s ou s t p erformance 7 5 5 Bui ld inq—in Resp onse In many circumstances it would be more realistic if the Data Acqu i s iti on and Reduc ti on se gmen t responded to some of the user interactive a-a -a -essages sent it In this way t h e s egme nt woul d behave more like a n actual remote machine following method This can be accompl ished by the As well as the information currently in the User T r a f f ic T a b l e values are kept for the generation of expected res po n ses These values are al so loaded by the control program When the messages are sent from the Data benerator information is included in the messages to dictate the response required Whenever the Data A cqu is iti on and R e d u c t i o n segment r e c e i v e s a message it inspects this response fi eld If a response is required the segment crea tes a m es sag e in the r a -orma l fashion and transmits it It then trea ts the received messages in the norma l way m essage t h a t i s s e nt as a r e p l y w i l l it self g et reduced a t the remo te gateway Such messages contain a special that they are response messages is The flag to indicate N o t e t ha t t he re p ly is con t r o l l e d by t h e D a t a G e n era t or caisistent with making the Data Acquisition Red uction p a s s i v e as possi b l e so t ha t i t a cts in a slave m ode This segment as The response i n f o r m ati o n woul d h a ve t o i n c l u d e the delay t ime b e f o r e the response message length 7 6 Onl y one message is ever sent as a response Control of the Traffic Genera tor and the Experiment 7 6 1 The J unc t ion and S t r u c t u r e The t r a f f i c in ARPANET generator is designed to be controlled from any host Accordin gly any host may communica te with the Traffic Generator through a well defined interface ei ther through a resident controller or throu gh one resident elsewhere in the net which accepts ex ternal contro l and to maintain Ira- order to maintain the in tegrity of this in terface a clea r separation of the control functions from the measurement functions they mig ht otherwise affect it is better tha t a- thor ’ c h o controller resident in a gateway Tue c a- -ntro l pathways are shown in Fi g 7 3 Overall direction is i n v e s t e d in an e x p e r i m e n t a l h o s t s i t t i n g o u t s i d e SATNET in w1a - Lca - a - a controlling process t h e ‘c o n t r o l l er ’ r esides Th is p r o c e s s a c c e p t s commands i n t e r a c t i v e l y or f r o m a p r e v i o u s l y creato l command file validates them and passes them to the net Ia-’ he reverse direction it accepts messages from t h e net and acts on t Ja-em where necessary Corrmunication with t h e n e t is via one of the partici pating the ‘ serv ice ga t e w a y ’ - a- t o wa s as a relay involved wh ich f o r co nt rol purposes acts All control messages will be sent to the gateways The basic control functions at gateway level are - to i n i t i a t e an e x p e r i m e n t - to s t a r t and stop t r a f f i c q e n e r a t i o n - to c h a n g e t r a f f i c d i s t r i b u t i c n or s i m u l a t e p r o t o c o l p a r ar r et er s 7 6 2 T on t r o l l e r —sp e c i f i c F e a t u r e s Thu traffic wh i ch w i l l generator is controlled by a set of be common to all controllers s i m p l e commands The controller can be as sop histicated as the experimenter requires and the following a-na -t h u discussion describes some features of the UCL implementation The controller will accept comnianJs interactively However it a-i ill also execute command files which have been previously pre- pared for it and are resident in the same host The detailed syntax cf the various commands will await implementation but it shoul’a-l be p o s s a - b l e to n e s t files a nd the f o l l o w i n g local f i le c ommmands at l e a s t sho u ld be available RLALO f lename WAIT -interval ’ before executing next command LOOP co unter repeat RE TUI N 7 6 3 from this file level n times Gateway— leve l Functions The broad categories listed above are all functions requirin g - - ‘ _a- a-a- - a-a-- — -a- - a- - - — ---a-- — - — --- ’ -- - 8 -a-a-— -‘a - ’ -‘ — -a -— - - C ON TROLL ER DATA a- SERVICE GAl F A ’i’ ______ TEL TE _ _ GATEWA Y DATA N E T T _ GA TEWA Y DA A L_ _ _ _ T a- I _ J a’ I a ’ I I L _ _ _ _ _ _ a- a ’ l FIGURE 7 3 CONTROL PATHS IN THE a- ’a- I ’ E L L I TE NET - - ‘ a- r AD_ AUN IS 1119 LRICLA bIFIED LFIIVERSITY COLt LONDON ENGLAND DEPT Off STATISTIC5 ——ETC F S 17 3 UNIVERSITY COLLEGE LONDON APPANET PROJECT U FEB 77 P T KIRSTEIN NUOOIlI—7lI C O25 TR—3 14 NL PU ci I S A 10 __________ 25 28 I2 2 F25 NATI ONAL MICNOCGPV — qcj BUR E AU O F S TANDARDS P FS O L J 7 0 8 ‘Is’ c ’ — O t h e t N r 3nd mos u n i v c i t y u sr% q Occu r 1r t h e m o r n i ng or a f te r n o o n — b ec au s of u n i v P r i t y l e c tu r e s ian f -86 - level commands w i l l lenerator t this be provided as packets arriving on performance by the gateway traffic a particular command port linking the gateway to the serv ice gateway or in the case of the latter to the controller through ARPANET listed below 1 The type of commands are Initialisation Setting up an experiment will involve three precesses The first is to establish the controller-serv ce ga te— way relationship A request to control an experiment is made to a listening well-known port in the gateway Each gateway in the experiment has such a port and if such a request is accepted the other gateways are informed and instructed to refuse requests until the exper iment is completed Similar command connections between the service gateway and the others would then in this w a y t h e w h o l e n et can be dedicated be set up to t h e c o n t r o l l e r e x c e p t f o r NCC i n t e r v e n t i o n T he f i n a l s t a g e is t o g i v e e ach g a t e w a y t h e i n i t i a l i s a t i o n d a t a t h e y need to g e n e r a te t r a f f i c and c o l l e c t d at a These s e c t i o n s w i l l be i d e n t i c a l to p a r a m e t e r change messages - accordingly t h e y d i l l be d i scussed in more detail there it Traffic Control The decision to commence traffic generation and to cease it comes from the controller in an explicit command gene ra ted iii This command app lies to all traffic being Parameter Change The parameter change mechanism can oniy be described in general terms The significance of the particular values w ill depend entirely on the traffic type being simulated The Change Packet must contain _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - — — t 1 - i 1 r n e i •r c f o t he l f l i c nt i i va I ac pa r i t e r i n f o r m a t i o n n a T r a t i c De e p t i o n or r e comp l e te l y T h e pa ek t c e a s he a for ‘ hr r — tie r r c r 1 r — tIe t irt f i r oi ch er eel c ’ o f IT 1CS r t r e I te — t I ’ ’ v r r -- the — h 0 e L i 1 S at ‘ rae r i ca a nt - ev l at sr i c e I T J r a t r n f - cs r i ’ i• - J i t o i l I f l t or toe r i H ti V _ i I l S S i •_i r ’ r’ nI C u l it i I from S ucceSi-i a t h u Such tCi r e ii j J h C C — V M e s s i qe s t _ 1’ re 1 — r ’ zr f i ed iO f lS d a ’ ir e l Len i t d t i r e ion role e T ‘h i o u I at ed r es c r • V tc- c ’ c r n ec t — — • u e e ed ‘r t he iv can be sped P eV — r s to be c r r _ c code i J The — flU fl1 T e ter ch p ari d —the r ct Jan • uu T ’ - t I n et tha corri r 11cr w i l l t a I l e r e e n ir I J h t in c i n h — tt i — no — p r t t ocr i — in — ic being tr uii f t i i i S L Ot be rru q - tu a ’ l h a c ’ O r t ij j i c r r ’e r ’ t 1 r ruI se n t i i ar n et ’r c h r n p crI e e i sa I i T r s u c c es u 1 i ’- r-o pro I r s r e U ’ -c eS r ht b ltCc j exr ’eriocri t a i r e i i i I -I o1 pI p -8 8- 7 7 rate of- 1 ha -1easurer ’xcr t Taoi S the Controller and Traffic Generator have A lcst version af eOn written ar J comI isS ianed h at h Both are s t i l l u n d e r g o i n g i r a d if t e a t i o n s as our experience with their use increases Some u r o qr a r n s have already been wr it t en to collect t h e d a t a f r o m the h r-r e the gateway software arrived from BBN only during December e x p e r i l i l er it s a l t h o ug h t h e r e is s t i l l much to be done in t h i s a r e a 1i 7 ear experience wit h our too1 The f i r s t a nd g at e w is v e ry I ir IL d t a s k has been to calibrat e t h e perform an ce f o u r t e s t t h e g e n e rat i o n and d a t a c o l l e c t u n as J t s of t h e For t h i s we c ir r a n q e a to h a v e t e e ge n i I a t or t r a n s m i t Sy s t i e r me s sag e s to i t s own g a t e w a y a c tu a l ly t r a n s m i t t e d i n to ‘ I ni s n o un s t h a t t I e n c tw t k ii cr y messages arc nut t he ar e t u r ne d a r e r n a’ a l aw l e v e l i n the g a t e w a y an d -ire handed t o t e e Da t a R c d u c t i e n F i g u r e s 7 4 7 5 an u 7 b dcnv n s t r a te sane t r J l c a i r e - i a lt S P d u 1e I n each ca s e over 1000 m e s s a g e s w e used to b u i l d t e e h i s t o g r ams shown ‘hese ha stugram s are d anl- tratc pr c ae n t uir t t ar tut ore e a ’or inents the JJC rt crn nce we e u ni t ’ x ’ ’ r ’t in o u r q i int e rnally mp rt i-p od This s hiaher ir u v e st iu a t e d it t ew ay F r f ou r h r v e sn ap ied ta r e J u c e • tI o sa i tw a r e ar t r i de l at of an be seen ta be ab out 48ms ann we expected and the reaseran mer i c n t as thor i n s ta n c e t h e x au ral interactive user c a r th e for this are being BI3N expects to modify the internal tue software they processing t rre of the gateway ihe App lication of the Measurement Tools Planned for SATNET 7 d It e i r be • een from Fi g 7 3 that the gateways are at the c x o e r i t i e s of SATNF’I’ They also may be viewed a dA TNET hosts or T h e y a r e t I F O fax c an excc 11— o ’ Jri a SA’i h I- F t e r I r n I mu i t ip 1 c’xors er t p l i e ta be d o i n g r iui surem enta on all of t h e pros t i l e modes ot use of SAThF T It blue edge of t i e Normally traffic will ori g i na te it one p o i n t network designated or a single destination at iri th r ‘l h r usual me asurement mode will be ta generate traffic at ne qateda y and collect it at anoth er As experimentation rtterr w i l b# extended to generating multiple s t r e a m s of tr a ffic bound for different destinations and eventually proceeds th ts V - — ‘ ‘ g en be ‘id e reecl fl c u r r’ utiur V r a - ‘O L -i t - i ’ — r f f C e cc’n c I i s c e r a g r a f c cu r n e r a - J s ‘oi L t hV s 1 c i l e c t i o di r e-ri to it h i - ’h e r u V Ia- hr - ut h t t ut - i 1 t o - f t i e t in eX i St S e u- i tIit - J - n at ia- u - c • u i t in - •’ d y e rr a St al i t - - CO r c a Ii -r C’ c - r l ’ i’ c o n di t n’ n t o u n t n i l aria i r o st - s r tutu Pu I -i J S Cn then i to tn s h ei r t r a V 1t ’ ro - - ‘- a V Si ’ ’ u’ i - - j i r 1 j i t e e i st i l l • - a - diii On x f ar x -- a even e r - e r - ’ a • in1 m i s t o b- — JOtLI so ’ v ad s t • C c s c c n- e rc dier a _ - 0 O- 11’ ’ ‘0 tO on tr ferp x— - r o e i cx h 0 - ’ a rue a -a a i t O or - n 7 7 at ‘i l c t a t -’ Lu t I o t I u n r t T II - Or - j FT SIPJ 0 p r ay I t h e a n gu ’u o n a n l - i iu nel b a u ’ io r f o let lilt - - t e r t - ’ an d i n c ‘t Flo r a o c r ’o at O t S -u ii vV - t ot LC- - i O 5 t U S S a OF ’ -0 i- - 1 e - vri i l oh n -OCj I r c l - t -t ci •i ut -t t aOe t -a s ’ e - V l a t r ru ’ oci - e e t i ’ JI iu pr u i” t i c rac e sire cr ’ r t a S L U ’ and e ili be r i e v a n t a dAT tb T the iaL i a o L w - f ’ tIe r C v a ie 1 V r t c-i a h l s a l t ir on p o st t ‘ - a d O V b r O air i s ta c s The t •il C-ia ’ u k i r i -e I - r i e r L ’ - in inf -a oi jr r o o ’ ‘ ‘ - ‘ai r i y 0 U I ’ th e b - asi c t r a fau c i c u o e i t it-S u i — er ’ a- su it ia •i fl rn it at t o in- - p i a r f e r l r u ’P rv t u a l 1 i on a lre r coo a rV c e s s a ly cn t 1a IL st co L- iou dA h i t a n s a n o r e on r -nt - -- -- ‘ V 7 t 1- en lair - - - - - - - - ----- ‘- - - - 8 p 4 4 4 u-i a -J a z D a I— 1 r -4- 1 • II L 2 Z _ _ _- - _ _ -_ - _ a - -_ —---- - _ - _ _ 0 1 D z LU _ _ _ _ _ _ -_ I— _ _ _ S - - 2 _ ‘ I L Z c o a uJ — -- - - - _ CD Lfl 11 a - — _ LU _ _ _ 3 _ _ I cj n_V to 0 LU C a ci -J -‘ S -- a _ -V J 2 N to HJ I— - ‘ —‘1 a g a IIJ I 1 I i ——- ‘ r ° -I j IN _ _ _ _ _ t - -j 1• 1 I - I - 8 J i1d -- — V _ -V —_ - - c3 -J 0 D 2 • U ‘4 -_ 4 13 C 4II a I ‘ I_ -V — 2 r V_V_i I ‘ I- C — VI I_ -i - ‘ S - 2 LU - 0 C’ — i ’ - -‘ 4 I r ‘ ‘ ‘a ‘4 — _ —-_ --- V - V _ - V _ V - - - - - _ _ _ -_ V _ _ _ _ _ t Li a CD I 4 f t E J - a C - Li g CD U i r d ci g I CD a EZ 5’ 1 -J - I -V_i § a • 2 ‘ V-i I — g ‘ 0 j 0 I 0 T 1 T 0 0 1 IN3JH - - _ _ _ _ I 0 1 3d _ _ 0 T_ V V 0 Vl 1 0 _ V - V y _ - • II — 0 1 IVV t j • S 1 - V - j V f l L O O d U C L L S ll In L i p i c ’ ’ ou a n n e a l n e r o 0 o f K i r a t i n 1 7 t A r e n or t n e t - r k mea surem ent t ools -ih i a we t egc t r cr with the programs needed ‘ or the d r ca - ‘-a- d e v ei c r - o ’ i rind i l y S i s of l l - T is be an g ‘1Iti - ‘ - i a - i - a d OiiO 3 U 7 7A f tj ’ a U ’ a l L ’ e ‘ revie chapter we give a r of cribe d full y elsewhere IL1L uC o c qU i S i t 3o t t h e d at a the signi c om p l e u r e r it p o t - c’ x t c d n 1 - r an of t yp e s i i i f G t ar t s u s ag e uri c t o and ha- e r ab l e -at a c c u r a t e s t at i s t i c s on V I t’ J o iiue to —_ r l v in - - r j ’ b t - elm or era the r in g rt - z i x l t e i -dl’ ‘j i r e s c i t e nd dli u oq’ S L l l ’O tj e the v er y t’ - o e r x i h c it a u i i c a u r fo r nt -- y i n n r n i n c r s and h is b ee n carr h a v e oPt r ned U e i - O j L ’ of u n RI 0 w tV j 11 ‘ to rI l e ‘ h ‘ - c i re r e I i l Pt a’o t i e’’ t’a u t l ‘ 01 4° 3 fl ’ 1 5 ar-a thr’ Re St •r e l- ’ CCV t iOt C s oo l y as ‘u r u L U- i l l no - ‘O itt t tx-r - ‘ ucban ’ s e a-r o t an c ’ct i o n S that - O 8O OLC C this lou i_ c t o ’ d i s c u s s ed blt ’ s c V 1 a - r c ’s s ar o -u or - ‘ ‘S t r - au’ -r a - ’ tar Cr -i 0- VJCttl’O STh 0 0 ’’ ‘ ri ’ - - -‘u d c — - ia h - r r t ‘ o t as of t t e -ant- ’et ’ K a i i - if at t- th - 00-IIC 1 a ’ n i l i- l a i t P 1fl t e r 1 a tV f ‘c - _ 8 2 ” ’ t h i s o a - I Y - ’ t t i r r ni q h Ph is v i a ‘ Phd eysteii p-c-n’t r t a r t s i l l i n t c r a c t 1 0 0 5 t o ‘ugh i t a L eat— The do a i s j r ri s i r - l i o n- C - a l ’ h u ’ c_ r fri ll datu on su It u i r c -J if f ’re n t n ’- n l f i c a ’l t ly aa -r s run for afternoons lit-I ‘ eat -J 1 tr- u i t a t aP c a a h - a r - -d -at Wc i u v e op e lr b t a s h by PI P 9 1 i u i n a rrI n x O V axch ii V t i r 1t ’ -a- I t r e J r i r e S e t h e r pres - ros an t i c o -t au F i r st i - i r s t in cr n r iy h i S i rt -a — ar t - V 1Pi t i a rr r Su 1t L ro - x e - - a t r a t r Va i t t ht’sr t o ‘7 C L a d e n i u p p o r t s on t h e L n a c r — ‘rIp an d i t id u i s p r o c - i 10ac c e s s c -an t r o l r i e c h a n i s r ’i rae i o t o il n a nt i - ti ‘eOc ’ at a t h e r i n g Th e i TI- r -s d t s e J e c r c d -ru c n r u t n u i n a resented in clue course h o w e v e r a I the ork a f l L t ot ’ e • - oar a ‘ c aa h i c -at Pave b ee-n d v 1 c - d further n ur t n h r C 1 S t ‘ a a r developm ent dark is now at n end and a i t r ix’ to the end al orting period - British o’ a nj n q l oan w h i c h th e u s er s tn f - in’ -itjon uti fri l I - ’ tic xu c V n V s 1 ’ ru - t that t- r icli oi -n m d ’ i ’cr ’ — s ci m -v h i’ r I ra i d e t h e V ’ I l ’fll’’ 1B iii o— r t s ir a b l e b uso I W I ’ t iac RL uiac hire a cc e s u xr - ic - c ’rin t -if — - — -‘_ ‘ — -— _ —-- V_ _ — — - _ -_ -_ - - We is i’ cc i i i r - ’-C” S-c - -us e d r - ’” ’r - l -c T’ C i V i_ r e i u ’rr r s ir la o - R P - N —_ -_ V- —-— -- --_- - — V _ V__I — ‘- -V — 94 — ‘1uch of the usage of RL through the TIP is from U1’ u s e r s m a i n i y local u s e r s w i t h h a r d w i r e d t e r m i n a l s or a few remote U 5 C 1 D w i t h leased lines In general i t is d i f f ic u l t to ob ta in a cons is ten t p i c t u r e of leased l i n e usage This i n f o r m a t i o n is v a l u a b l e f o r example to PTTs who wish to know whether the characteristics of lease d line usage differ significant ly f r o m those f o r PSTN u s a g e Our analyses do indeed show significant differences and the results are given in Section 8 3 The third measurement tool runs on PDP9B in conjunction with the m o n i t o r i n g and access control program QUES m e n t i o n e d above In g e n e r a l QUES o b t a i n s i t s i n f o r m a t i o n from the u s e r and then c l o s e s the connection provided the user is authorised and allows him to log in as usual However it is possible to make the connection for t h e u s e r and s i n c e now a l l d a t a passes t h r o u g h t h e PDP9B we a r e able t o make a comp lete transcript of the interaction Due to the h igh cpu overheads involved we have only used this technique with one host w ith users of the National Library of Medicine mation retrieval systems NLM infor— Analysis of the data thus obtained allows us to d e t e r m i n e various parameters of usage of such a s y s t e m f o r example the distribution of system keywords Although the tools have been developed com p letely in 19 7 6 the number of runs made have been very few However we plan a week long exper iment early i n 1 977 w h i c h w i l l provide a s i g n i f i c a n t q u a n t i t y of data Our activities in this area are recorded in Section 8 4 A n o t h e r m e a s u r e m e n t tool is t o t a l ly d i f f e r e n t f r o m t h o s e d e s c r i b e d above The p r e v i o u s t h r e e were a l l at t h e user l e v e l t h i s is at t h e line level i e measurement of the IMP subnet W h i l e there have been exten s ive measurements of t he s u bnet b y the N etwork M e a sureme nt Ce n t er NMC at UCLA these measurements have relied almos t entirel y on cumulative statts tic s gathered by the IMPs or messages traced through the subnet As a result of these measurements the NMC has identif ied a number of significant problems in the subnet par t ic- ularly logical errors in the IMP—IMP protocol However for cer t ain specific purposes in particular the TCP of Chapter 5 we have been interes ted in measuring the IMP-IMP traffic ourselves We have set up some f r l y elaborate tools to make thi s possible but they have b ee n used ord y a number of ‘ s p a - a i r Ic effect or t tr protocols 1 1 ’ is de s c ri b ed V r I Final l v c a r t e ac s p e c i fi c N et w- u k ‘-°ntr -l of bee n 1o O t rood - ntations 8 2 to ‘a N on it o rin t The t echni - ruues It a k V C r t ar -C with 8 5 eat o ’ i _ t i e out - - - o r- T h e s e r r c as u r em e n s O Ptu -O O d i i te rent at 8 i op turf - lflt aC’ — • to itor i - some de f ii i in toe ri-ar L Se c t t iV t c ’l ’ P a ir a im s 1 s - t r - s - -l V STN it e rr ‘ ‘ i r d s t i c k a f a r o a t ’ us art- ’ in S c ’r t i on r r o r n r t h a ve been iS a’CP it elI t - t ‘ r c te line on i_ Lu IMP— ’04P satell — t o i n f ‘T I - b e d u s q- h i - t- t b C - i t o ‘a t l e t pu 0 a rim — b ‘I ’hii s - ‘Lr r o i - ‘- ‘ a - e o - i r ’a Otl S innua l t ’ di N were descr L - c a 1 0 9 7 6 A a n l s az - c F i r s t e i n vii -o ’ t t simple result ‘ - re p i r - s u n t e d hit ’ e cf s x r ’rc e h a s b een ‘ad€ t o cft da t a ‘iatherin j ui -ngr- on h u ’ I t t e an a t y s I p r a nr a m s hav e o r modified and - o x t -ride I lx ’ a year ’ s riuoru itc — ri r - - l t - u t Fo r c l ar t v report V J C PVP 9 B V L QUES at it d - tn t luirar d r L h of ‘ia na ’a dt ’’°ani ic tat ’ a u o 1 for tue other it se ll 00W hat- c re sults f-a r t e n e t i t -n tin 1 arid in h e rr c -q at c s h e id e d t h e s e a r e a c c e p ta b l e i _ -- i nest nuxn € -r of 1 II pe r a o c al I c o p or t s e ur o— ‘ for his name not checked t h e h o s t he w i s h e s t to It t h e n a t t e m p t s at due remote host required ari d scr o n It QUEF one secon d i n t i r v a l s tC nne -t— then asks him t h e next host number For every interaction it pri Pts data -onto a paper to with other information e g every half hour it prints a m essage stating it is still monit oring T h i s paper tape is then cez t t- the RL 360 195 an d analysed 1 ’ ’ a program called XFSTATS Stokes _____ hi he user 2 i seconds to connect ac reconnect to the user -and onl y succeeds when h e closes the inn Pr- -v— ac ess is a l l o w e d one er r o r on e a c h closes the conne t i or a - illawe the required C ut e r ru -at are VJI V fl a u s e r d i a l s in t hui t h a s QU ES i r r V m e d i a t e l y seizes t h e t - 1e u ’ar T I P p a s s-a er d and t h e n o o h e r - t r e s u i t_ — - ‘ ‘i-a u ttrr - connect on Ut n o t i r i e -f u r m - o c - -s r u n s en t l y t ho s e a v a i l a b l e t h r c u o h t lie PSTN n re a k s the c a o i e c t i-c a V c h urt - t r use t no w ay T h i s p r a 4 ra0 c ’ n u e ct s t a-n We - - - — - IT-- - 19 7 6 B T h i s is w r i t t e n i n Babbage S t - r k e s 1 97 4 a i v e use of t h e f a c il i t i e s in i t s aystem I ib i o r y The analys is consists of two phases and makes exten— t h - I 9 7 7 1 The f i r s t i s to r e d u c e t h e output to a canonical form this consists of reducing each total i n t e r a c t i o n to a s i n g l e l i n e and t i d y i n g up in t e case of system crashes Considerable error checking is performed and the output a n a l si s in t h i s t h e d a t a is read into core in a specified structure may be assumed to be e r r o r f r e e The second p h a se is t h e a c t u a l and this structure is traversed in the requisite fashion m a n n e r we a r e a b l e i n some - J e t o i l to a na ly s e u s age of In this t h e L o n d o n TIP v i a t h e PSTN To he s p e c i f i c t h e a n a l y s e s a v a i l a b l e a r e T imes mon i tore d iclent Surnames used for each Number of logins connect i_ joe and number of hosts used for each Number of Ident l o g i n s c o n n e c t t i m e and number of users f o r each host M a t r i x of c on n e c t t i m e s lao each trost number ut--m t pair Pr o simp l ic it y the results are available in grap hical or tabular form We have a considerable amount of data for the period July 1975 to J u l y 1976 and it is upon this that we concentrate Before that da te although password validation was effective we did not prohibit access for illegal users After July 1976 we have done relat- ively little such monitoring due to other pressures on the computer Over the year in question we monitored the TIP for about 5O’ of the time but much of this time was at weekends and in the evenings when usage was extremely low In addition the -rnount of monitor inu performed varies significantly from month to month Figs 8 1 and 8 2 show t h i s v a r i a t i o n the former shows the breakdc-wn by h o u r of t h e latter by month day The largest usage was b’ t h e B r i t i s h L i b r a r y r e s e a r c h g r o u p s a c c e s s i n g t h e National Library of ‘ 1-o icine datab ases However this use is b i a s e d because the usage of NLM coincides a l m o s t e x a c t l y w i t h t h e t i m e s we m o n i t o r - -V- —‘ - -V-V ---- - - - -- - - - -- -- 7 i—a Fc UJ 1 70 4 H __ 50 — 1 i 10__ o - ____________ 14 0 ma ---- - - I -t - - ‘-- — - - - - ___ — - -- - -— - - - —- -- i2 i r i n -------V - ’-- ° r f e tn i 16 t-i I- tI1t- tI ’E t ori - 1’ m - - - OF t ’ — Zj 70 60 ’ 50 40’ 1 — 30 J U L 20 ’ 7 5 10 0 - A S U G E P 7 7 5 5 O N D C o T V F A N E B M A R A M P R J A Y U N 7 7 7 7 7 7 7 7 7 5 5 5 6 6 6 6 6 6 70°A 614Z 34% 142% 514% 51% 39% 37% 56% 36% 36% 147% MON 1- ig b 2 _ J E C QULS Monitor in q T H Per f r n i t - I - - - V - - - - - - - - — - the other band most ufliv -t sity usage occurs in the Out morning or afternoon - the more active u s e r s P hy s i c s U s e r s to the TIP 1 nn of Blacknest UcI for al U and the h igh E n e r jy because of univer oity lectures ate access v i a t h e RL 360 1 or a d i r e c t leased l i n e Thus their usage is n-a t rerlected i n t h e f i g u r e s The overall pattern of usage monitored ov e r the year I S sh- -’ or ir F i q s 8 3 and 8 4 The r e s u l t s a r e e x an i xuorm in d e t a i l in stokes 19 7 7 A Another type of statistic of interest to P-ITs rind s y s t e m s m an ag e r i s i_ h It of port uscue the F i g u r e 8 5 chews f o r e x u I c the number of logins hour monit ored during specific periods of i_ tm day 8 3 Measurement of Usage via PDP9A PDP9A frcntends a number of - ou lpu ters onto ARPANET as has been discussed in Section 2 and the RSRE 4080 All traffic through PDP9A traffic a r e t w o ty pes of s - l a r to t h a t l i n e s to t h e ‘rip A t the present time these are toe RL 360 195 mat t iC’ -QU Pf a is monitored There hence of mea sureme nt o The ii st is Section 8 2 instead of mor itorinct f r o m the P STN we m oni h-o r t e r m i n a ls ot t - a c h e d the RL 360 either ritreu tly or ‘ ‘ia a w o r k s t a ti o n is the converse the u rmuqo tc The second u s e r e A P P NI ’I accessing the RL 360 or f r r rr - RSRE 4080 as a host The a n o i ycec performed are similar to t h o s e f o r QUES a l t h o u ch in a d d t tm r a n system performance is monitored For users to A RPANE through the PDP9 the analysi available is identical to that fet- QIi ES s i n c e we a r e c o n s i d e r i n g e s se n t i a l l y t h e same i n t e r a c t i o n s t hat is a u s e r connected to a h o s t on t h e London IMP in one case the TIP in f-he other PDP9A with the sing le exception that the user is identi fied merely by h i s terminal address For usage from A RPAN ET again similar d a t a is noted as in Section 8 2 now however the source host and 360 wa-i IDENT are noted ir stead of the destination host and u s or p a s s w o r d of S e c t i o n 8 On a s y s t e m p e r f o r m a n c e the information collected is the time at which each machine JUP RL 36 0 and R SRE came up again 4080 en t down PDP9A s c h e d u l e d and unscheduled -and Each occasion of receipt of an incorrect protocc- l m essage is noted In a d d i t i o n the start time end sourc e and des- -- -- _ - - — BRUNEL U Q M C 2% — I 1% RFADIN GU 27 R C 100 OTH ER 10% Cr o B U R G H BRIT ISH TH A iES PO LYTE CHN I 1 8 3 Usage 0i 5 4% y - PETER KIRS TE L IBRAR C 8% IN AI 1’ANET by V a r i ou s R e s e a r ch G ro u p s NAsA-N Es 1 3% CIIU-A 1 5% OTHER 1 6% MULTICS 9% SRI—A UTAH 4% 4% MIT-AL 5% NATI ONAL LIBRARY O F MEDI C I N E - B B N 5 1% 6% I S 15% Fig 8 4 - Usa ge of Various ARPANET Hosts - - - -V - - - - V NO OF LOG INS 3- 2- 1W EEKENDS 0 — 0 V VF 6 12 TIME Fig 8 5 N ura b er of 214 18 OF DAY Loqins per h our - V V •V - ---- — - - 103 — tination of each file tran sfc —V jS -V _ _ _ _ _ _ also noted From the performance data noted we derive igures for th - Mean Tim e Between Failure Repair MTBF and the M ea n T ime to MTTR and the frequency of erroneous pro tocol messages and their source From the data collected on ARPANET usage from the RL 360 we derive r e s u l t s ider tcal to those of Section 8 2 with terminal 1 cation rep lacing user ID In many cases this is more r e l i a b l e because a aroup termina l is used exclusively by a specific From t h e data collected o L 3 0 ard RSRE usage ag ain the resul ts der ived are identical in form to those of Section 8 2 now toe service host and RL 360 ID are used as identifiers The For files trans- file transfers are analysed in some detail fers between the PDP9 and ARPANET w e list the number of t l l s by the Ident of the person initiating them and t he elapsed time and direction of each file transfer Finally histograr- s are der ived of port usage for each machine the distribution of port usage as a func t ion of t ime i s pre s en te d On examining th e g lobal statistics over 1976 w e find the availability of all the mach i nes is h i g h times the PDP9 crashed tha t The number of 847 is also high but on examining thc data more closely it can be seen that on 5 occasions a single fault e g a host on ARPANET sending or a drum fault on PDP 9A manifested itself illegal messages and caused th e If this occurred at n qht with no—one pres e n t the auto—restart mechanism would enable the crash to recur system to crash at intervals of a few minutes until correc ted t hus giving rise to a high number of crashes 307 crashes were due to t his and hence the rea l r umber should be 545 giving of over 9 hours Another global statis tic is the num ber were in use a PTBF of times all ports PDP 9A has five ports to ARPANET and on many occasions 155 times during the year t hese were fully occupied From PDP9a to the RL machine t here are o n l y t h r e e ports available to ARPANET users two other po r t s are ava i lable the first is permanentl y connected to a specifi r - - - - - --- - j _ _ - -V _ -V local device and t he second i s u sed for t r a f i i c in the other direction The number of t ime s a fou rt h user tried to login and was rejec ted was 358 Fi g 8 6 shows the total numbe r of logins to the PDP9A dur ing 1976 On examining the connec tion statistics Fig 8 7 we see that the greater part of usage of the RL 36 0 is made by the UCL research group INDRA and the Blacknest Seismic Research Centre who have set up a seismic da tabase which is updated daily The relatively small amount of usage by other groups is a reflection of the nature of host-to— host use of the network such use tends to be concentrated and concerned with the routine running of jobs or the exchange of files compared with dialled-up terminal access which is often concerned with communication and exploratory work Host—to—host access ther efore tends to be of much shorter duration than terminal access 8 4 Characteristics of a Specific Host A par t of our project has been to provide access to a number of British Library BL centres accessing the ME DLINE and CA NCERLINE databases on the NLM IBM 370s and to monitor such access Some moni toring can be performed by QUES other monitoring is done by a program called MEDLINE which runs in conjunc tion with QUES If a user specifies his requ i red host as 1 47 the TIP to which NLM is connected or NLM he is not released a s usually happens with QUES but rather is transferred to M EDLINE Thi s program maintains status information about NLM and informs the user of the curren t status If satisfactory i t is assume d he wishes to proceed- if not he is asked whether he w ishes to wait while a connection is attempted The program attempts to connec t to the f i ve ports in turn and if it is able to do so logs the user in using either an Iden t specified by the user or a free one chosen by the program out of the group allocated to UK users — 105 — AV NUMBER OF LOGINS PER UP J J 1 J JJJ J J 1 0 0 5 0 - — — 0 Fig 8 6 4 8 12 16 _ 20 214 _ _ TI1 E OF fl V Numbe r of Logins Per Hour on PDP9A - --—-- -- - - - — -— ---- -- — — —-4 - - 106 - OTHER 17 0% - - Th “GUEST” 7 1% INDRA HIGH ENERGY PHYSICS 6 8% Ji ’ IZ ’ 1L J O SEISMI C ‘ LU Fig 8 7 Usage of RL 360 195 by Var ious Groups Via UCL PDP9A - 107 - It then asks the users for h i s nassword and from this staqe on permits a transparent connection with a few exce otions These exceotions cover three areas The first is that simple control characters are perm itted to allow slight local editinci of the current inout line the second is that the user is given access to certain data held in the PDP9 e g the total connect time of the current interaction the third is that some of the verbosity of the NLM dialogue is removed Up to this point the program merely acts as a Network Access Machine Hcwever if the monit oring option is selected a or more specifically direct transcript of the conversation a transcript of all conversations in progres s datest axnped and with a channe l identifier multiplexed together is sent to the current monitoring device usually magnetic tape but disk if required This tape is then sent to the RL 360 for analysis The analysis consists of two phases the first is to d mu 1tiplex t ’ie separate conversations the second is to examine chara teristics of t e usage for example occuriences of system keywords At the oresent time we have considerable data on NLM usaue from OUES but little from the MEDLINE program due to various technical oroblems A week long experiment is planned in early February 1977 and the results from this will be reported later The results obtained from OUES provide some interestincl insiahts into usage of NLM Because of the mode of operatiofl of QUES some connections are recorded when in fact no real connection was made This is due to a connection being made to the appropriate port on the NBS-TIP correctly but itself being down Rem oving these from the statistics and also unfortunately some very short genuine interactions we find that the average time per session is just under 23 minute s see Fig 8 7 indicating that moSt users perform one search per session in one session however NLM - thirteen s’ arc ies were performed in nearly two hours involving the outpi t of over 20 OO characters from NL M 8 5 8 5 1 NetworK—Level Performance Measurement Introduction In prev ious years we have constructed simple measurenent tools - k -V - - - - - t o lo o k a t p a r t i c u l a r t h r o u g h p u t d e l a y c h ur a c t e r i s t i c s f o r traffic be tween the PDP 9s and o ther A RPA N ET h o st s or to look a t p a r t i c 1 i it provem ents in the ARPANET N e t w o r k C o n t r o l P r o g r a m N C that we were thinking of making We found usually that the effort to get reasonable measurements was not negligible and moreover we w ere l e f t w i t h a d a t a r e d u c t i o n p r o b l e m to solve in each case Unlike t h e TCP e v a l u a t i o n e f f o r t of C h a p t e r 5 we were n o t prepared to b u i l d p o w e r f u l gen eral measurement tools as the se t of m e a s u r e m e n t s we wan ted to make was a changing requirement from time to time The i n t r o d u c t i o n of S%JI TCII C h a p t e r 3 was a good o p p o r t u n i ty f o r des ign in g a th roughput d e l ay tool that could b e used e q u a l l y to loo k a t th r o u g h pu t d e l a y to dest in a t ions or at a particular NC performance inok The SWITCH i n t e r f a c e a l l o w e d t h e t r a f f i c g e n e r a t o r to l i k e a ne twork itself and to be accessible for running exper- iments as a similar interface It was now justifiable to put effort i n t o c o n s t r u c t i n g automatic generation collec ti on reduc tion and g r a p h i n g f a c i l i t i e s i n t o t h e tool b e c a u s e of C er t a in l y i t s g e n e r a l i t y of use in our contex t the usefulness of this measurement f a c i l it y is ak i n to a p e r f o r m ance m eas u remen t tool i n an opera ti ng system We wish to evaluate possible improvements in our network support software to monitor changes in subnet protocols or topology over a period of time in order to detect any effects which m a y not have been antic ipated Since in our system we are s u p p o r t i n g host connections by RJL-type protocols we w ish to gathe r perfor- nonc e f i l u r i - - on these hosts par ti cul ar l y where in the ULCC - i t t a c h m e n t we a r e u s i ng h a l f — d u p l e x p r o t o c o l s w i t h low t h r o u g h p u t potential In the case of EPSS this tool may be useful in provid- ing thorough statistics on the switching exchange availability from a user point of view A facility of this type for performance measurement is a useful additional tool for the design of a virtual circuit mapping gateway A full descriptIon of this measurement tool is given elsewhere Williams 1976 here we summarise the tool facilities and some preliminary measurements 8 5 2 Tool Facilities A traffic generator capable of varying frequency and length of i_V - -V - -V-— - — - - - — - - --_ - VV V -_ _ _ _ _ _ _ V --- -_ _ _ _ _ - - -- — ---________ kes- age gene r ticn resides on t h e h W i F C H g A Le - y as c network- address It is c i b I 1 of - eri ra 1 ri to n o h e r of o e F e ir a t 1o r s or n f l a number of onnect i u i s I n the s-in c d e s t i r a t ion sir ru l r- ir— eously Traffic 1s timestamped on generati m ard on r e c e i p t of echoed messages allowing o v e r a l l u e i a y round—trip time Additional leve l when necess urv t o be ‘ e a s ur e d hr h lv i n c t im estamps non he inserted at toc N i The m e t h o d does requir c an echo p r o ce s s t c he used at the destina t i on for hosts i t i s p o s s i b l e t o g e t r o u n d this by using a proton positive a c k n o wl e d g e m e n t delay sidraJ to me su r c If necessary an echo process can be - -irit tea at the des tin - ation to allow additional t i m e s t a rn p i n g The r e t u r n e d m e s s a g e s have t i m e st o i p s r er uvcc3 a n d his ’ o r y records are written to magnetic tape in an IBM-compatible format An o f fl i n e an a 1 y s i s p r o g r a m t a k - s t h e r e c cr c i s and f o r m a t s them suitabl y for inpu t to t h e PMD g r ap h i n g p a c k a g e a UCLA st a t i s t i c s package corr n o n ly a v t i L O L I C on 3 0s 8 5 3 Measurement Results At the tin e that the measu er not tools were completed o n l y t h e ULCC CDC 6 4 0 0 cnn e c ti o n was f u n c t i o n i n g to t h e E W ITCH interface t h e o t h e r h o s t s and n e t w o r k c on n e c t i o n s to SW I T C H w e r e s ti l l u n d e r development The CDC 6 4 0 0 proved a u s e f u l and of t h e tool as t h e p r o b l e m s -o f i n t e r o s t in o f i r s t application achieving reasonable throughput with half—dup lex protocols are we 1 i-known and we w i s h e d to c o n f i rm this behaviour on our own connection We also wished to examine the polling fre-luen cy and message processiro delay at ULCC as -iell as exdiii ri J fl’ h e b e h a v i o u r of our software which h a d n o t previously bc- en used in other than a test setup Measurement consisted of generating messages at fixed intervals to be sent to ULCC As we had no echo program available we used t h c protoco l acknow l edgement as an indication of delay to an ci from the CDC 6400 Uc isurem ents were carried out both at peak times and at night to facilitate control over measu rement periods L _ - _ _V_ __ _V_ _ _ _ _ — -V— V — LOC i SO v -n in Job t L t t — t e - 1 0 i- 1 aio - o t l c s hu t d o w n expe r i m o i t - l r u n - 0 res l s s ow d as t O i l L w s — t o a l o - r e 1 - i n u - S 110 m e a n t n c r o r a 1nc o sJ ” t o CDC 4 0 0 i t s k s - v L e O - c nt i i i i n t a ir l y c ar sistent pattern of t h r o u q h pu t e time e itJ I time in the C b C l D i - safl • - rc 2 32 secs 0 SOC $ 7 31 suns 0 S n r ct i nu t n e t h ir u f i g u e i coo oc con d i r - u w s a c o ns i d e rab l e t l i n e sp e Lt in no PD1’ o- j t s t I e ‘ li e a ctua l prot - ci drivcr for the 200UT e m u l a t i o n A l t h o i i ih in TV get buffered t o ba rk ri s t o r e t h o do - ene able t o t n f o r most O-c PDP 9 system messages do tune cs obviousl y ex cessive Ie’ in a o c h e du l ng er r o r ihi -n h ac c o u n t e d this d - I i ’ t h e PDP9 nr c -cesso- s are al 1 currently Qi cL o c r — d r i v n n Ti n e r e s o n u l e i ns a f f i c i e i i t - — in t h e n r c -m ind a c c u ra cy t no to tn d l Se of handler at o ur 0 0 1 5 secs w o u l d mea su rc-i’cnt c l o c k s is h i s f t g r n more precisely J O L t r i — 2 t i nq t h e t u - n d f ig u r e f ro m t h e f i rs t a ir e s t h e time outside the PDP9 as i h O c c O The 24 b y t e t e s t m e s s a g e s u s e d for these measuremen ts or e consi erobiv swollen by addressing m in or and p cid tnq c hs ranters the CDC p u t s o u t such things uS screen c i c a r - lor ac 4- -OTO and the c o l c u l at e i transmission n addition or mcs a- -S both directions is 0 51 secs there is a m a -n w i t ch 1 ng t ime OSsucidted with the use of a tir e protocol h a l f —du p l ex appear to i A l th c - igh the i r - u qh - i t n - r L n e e t I n n - to f g ur e s tb i m p l e m e nt a t i o n d ru p e e d f l a d O - i C in th e 6400 would therefore i n e x c e s s of one second per m e s s a g e n n o s ur e n s - nt s as h i g h as 7 3 7 l e uk S this s ligh t l y - T i m e sp e n t w e r e v e r y c un s i s t e n t o c ca s i o n a l del ay c-cs we r e r e c o r d e d i e ‘hiOO c a l c ul a t i d Tn by e m u la t i n g p in f or b o t h Si n g l e a n d connection is triple t c h i cv e d in t h r e e 20 0U T t e r m i n al s in m u l t i — T he CDC pol Is i n f r e q u e n t - ly a d d r e s s e s min t c u r r e n ti v jve so t h e a d d i t i o n a l o ve r h e a d 0f u s i n g t h e m u l t i - d r o p p e d software is mar g in -mi Iowever it is interesting to compare wh et h e r the sum of L r e e i iademl c o n n e c t i o n s compares w i t h throughput on 1 s i n gl e m r e c t- io n the - L V _ ‘li e i ps c_n ‘‘ l i L L i 375 bps a nhir -uds mt u t-a ‘ a 1m - ti Tb i a i tim ‘ m — - i’ 1f a io r m n i u c i - n - c a n i-m he f t i m th e - i ins t r i- m - ‘n o j -t t ’ c - no t - -i slit- aid Link err I to t a l k ifitl at I rm- j i - _ - m m - ’ ir il i - i it -I lit mOt - ’ r le i V - - m • V V _ f r h u t - — - - ci -a h 11it O c m i r m - m I - t t a c t a-i r - Is k i nd 1 n - - c -- ‘ t U -t’ - i - s I ni--ra te U- a s r ci c’ c ” - -r ‘ I o t qn s Lin m- om C’if S — t - 0 ‘c it -l I Ii- - - ci _ -’ ’r m- ‘ 0 i3 m l I 1 -1 hr ’c-a - im t i me mr n _ s_ t en u a l i L - - snvc- r moi ’’in - ech o r o c ke t s tim- i i 0 1 ont coma P nasurc ’r t r ot - l - - alS ouna i t n -‘- iC- - h l $ nh 0 o I- ’ ol al 1 i t s -cci pe m I I e l I I- - - ’ 0 l - CC s- - -j Y U - i_O h tOO-C cc m i n t - c t i n 1_ l i O f i i_ a un -ed ic ’ fl L - mc - - - th - r m- f m r c f l c use T 105105 can e nc n - r u i r e d i t I’ - SO 11- ohm f_ - 1 —s- t - U i t n — a t f nO 0 m- - m i I ’ ‘1’ ‘ t miS r - i c- ri tz -i ‘rm 0 c - nn l c n c - _ in j lj n -mCi t - n t e - diI t lit r n - c ’r a -p ‘ - - t ic- data - 1 rclIrmma i - Ii it -‘i rc- uuh out o- ’r s c -C 5 1 - P - rt cm m - jt f -n It - r si eC c-- m lr_ 1 - i C - -or - 0 - r- -r - 1 St - u a’ — d c c o n t i ’1 i- nm L o l - ii - _ i r s n n - A l n E - i -i-n - - ’ I O L ’ i t o - - l i i i ‘ - -a l miii u ‘ at f - aS i - r a - m a u n - i n n i _ s tIlOSm -i I - - Tum - t c o ot ‘ j I 0 mOO - in n - - - r a nt a ol L 1mm - mL ’ - r ri - -not d o — t j am 1 -s i i r r i 1-- n- ou - r- t nlr H - mfl - nc ‘ to - 15 c o r b i r e s - - xt2m 5 ‘u mT a r nw i r - ’ - in O C f c- - - 0 L t e r i ’ r o Unite o s s a o- f - -t - i c- x -m tm-ca - I i - lt c- ’ -‘ ‘ t t Ov these if v h l or - - o - m t ’ n I- ’ - - i 3- - i m r i t ui o n o h i n n ’ - irn rs ir ’ - I i- to r - t O OV s u r p n i s r tIi 3 f I o lu i l - - C -h’ T n _V- _- _ r P r n n r - m _ t ‘ r nr- nu’ n i t u r e a t fine ’ i P n - -’ n- ‘ n t L t u h i - C ’ i i J -5L17t j VV_ — - — - f- - rim a z ix - - — t c s m m 3m i-i one i a tom I l ’ I i at - ’ - l V in ‘‘ - -a p i a — n i t ‘ t - -f i--n o- n u n i ’c ’ r m n c1 c-- ri o f I - ‘ 1 1 1 JSc -_ Cm c m t I at mm i t i l l S al t 1 OHi a I 0 1 - i i—p - 0t ‘1’ 1 on c - it lIt 1 1 - r n o iP- ii f rets t mO V t J tO I- SOt ‘ t dc cc- mu i - t L t 0 ’ i I i c o lt d-11 -- ‘ I i- r r- a i Tli - i i l t - emi an t i rO of 0 - v - - o- a to i u_ i IL— H-C co u Lm C - - - il t 00 I mc L O p a ct - L ’- i i - h - --I C - 1mm ’ - a I t n i ’ ’ I un 5 i CO 0- ic 0 cI t im itit etO 1 0 cc 0 tm a- I’ - n e - r c -t o n o i of mc- - Li ’ cle ne I m b 1 t - - r J o i Yost of i _ n c t i c c-f - dat a 011- -u tIne hor t --d- st 0 e hu t- 1100 ’ I O rk I 01 - s - -u- - i_ f o r or Oa t a 0 n i ‘ 1 ‘ -t 1 t -fl ‘ -‘ ur m c ti ‘0cm - a - n i in for m - t 1 ‘ p - 0 0 1 0 t cf s o r ai i it u - c ‘ to so -m c-- - ’ - - -aii 0000 1’ t - u 1 _ C i ma r i - -c - c- - ao -O nil I - lomsO u i- - a c 1’ - i -t u i tm - ’ NI -i r o - - - 0 - CC - I aL -S a c- o i’ 1 0 0 1 1 i L - t t I ora te 1 0- 0 i o-Iv a 1 ot t e a’ -t nme’ i y t ou r 0- It Cc ‘- 0 -n - n i ci fln t- - - tic mu - - t t O ’ I n n - m o p e mi _ t 005 1 0 t i ta o ant tho-o- - uI — 1 00 cr i so - C um ’ - ‘‘ n s i d c c i t i no I or I -3 u - 1 - mr th rum -n - L I — fE i a n n i v - - a - -o n u’ - to C - C O t- S -i-_tO Y - mm - -V 1 1 1 6 5 UC 0 m U s ‘- ‘ _ -- o - - - ‘tI c- heacers aie - mrIou Lu uO t 0 aa I -- o t-n C 1 - - - - Pt V i I i-O i t i y I - it TO n -n - L - ui-L- -I c a 1 5 0 - £ r’ - ri Itt L-t - m m - - t s on ‘010 n- U m u r t e n t o- oi - t t e u ’1 am tm m u c t u 5 - C a IC i I a o - l y IT O b r -n -i- P - - a - —-- m ’ 0 ’ - - oc- r c i O l i O tO n - ’ - - - e’ I t I ’ t i ixI - 0 0 - nOt 0 ser i o us ‘ 10 - -0 - a toerrI Ly ym i L h I ‘ - tL-n and I- f Ii tUe out - 0 - - p V -_ — — - - --— - - - - -V - 5 a -a ’ o t m - am d D s D o t - a CuT O u - t 2 5 i - o - L ’ I o-c - mm t - - 0 en 1 t t i e a n a l - m a - r c m r m i f l j y m n m l s i s so- ssxu- n I n a t I - i - ml p r o C m i s w ir e wm L t i 0- - m It - - or ‘LUt - i T S t S c - m I S l i mO I cipt- and pr c i s c - e s m o m ’ - j r mam- m c c c a as - LublO of - -a-i i - - s o t - H o L im e 1 p0 imm 1’ m i m p oi - -o in f -c - x tmm it r u n C t L t O 1 1 m t ” ’ m - lit -f - - - C l iar tt ar -u - n i - I u ’ 1 77 T 1t _ _ _ _ _ - - -a ch Lb - i f l i t ’ I Lr t_ pa ck - -t on ic s m- ’ - im - ‘ Ill I ri 1 1 ’’ e n iin ’ - - - L tm- j atO- L of ‘it mc s m ’ - - - ‘u ii ef i c ’U tar p 1 tOt t ic the n-mi om - - m m m l n fei h i-w 3 i n i I - ‘ 1 5 c 1 T 0 ’ T i SittO-t -OIit OIt ut-i m - m ly F n ’s it n- c-f Fl - Iu t i t -- c-li h 1 f l d r-m i ’I’ i ’ O i I - a am i ’ - r t i d d- -t n t i moo mous in nit- i n s ’ dr - s o- - i b a a l t h ou g h V -V ti- VV - m tinO of Y ’r S i O f l V V - V --V - 13 may decide to omit c e r t a i n t y p e s of - - is o a c ket s u c h t c - I l r i -a packets These may be analysed separately tc i e t e r r o 1 n e f a n example the instantaneous state of the netwc— rk as seen tan the UCL IMP In add ition to the logs we are able to p r o d u c e five t U p e s a gr aph One exampli- a plot of the delay between the sendirm ct of a message and t h e R FNM in Fig 8 11 Ready for Next Message as 1lustrated This measurement tool has proved valuable on n u m b e r of a occasions Our preliminary results revealed t - -o i nterestin g as c- rs of the London—Norway line First the routing ne ssages accounted for over 60% of the total data traffic even when t h e l a t t e r i s high high Secondly the number of retr ansr iissic -n s - m - t o be nc- r T h i s is caused by t h e l a r g e a m o u n t s of ne isra ic dat i beinc injected at Norway about 2 4K bps aver ian a nd t i-me of packets that can be outstandincm at any unu of found t i n - c- these f a c t s with the long transit td e ovcr t h e circuit and the low bandwidth of the UCL-No rwa’ -and limited The rom - be r cc - binar r a ut e l l i t c Norway-Us l ines leads to the Norway IMP being sw a r ped anti reject c-ic data until it has freed some buffers This explained sorn e i t m c m n a i o u s results encountered with our TCP experirn c’nts see Secticn 5 3 2 when it was noted that contrary to expect itl c n s increasing the “ w indow ” size past a certain threshold a c t o m m l l y reduccd Another point of interest is that when ci h o s t i f l up n-Or it attempts to send “resets ” to e ver y other host or Due to t h e mode of operation of the protocols t h i s tllrou- tmmcut i MP c o m es the nctw ’ -n ’ results in the send ing of three times as many messages into t h e network and t h i s again causes the Norway IMP to flood T h u s this tool is a very valuable addition to our r e p e r t o i r e of mea surement tool and provides a flexible means of examining lowest level It is however most useful tho in areas where we have a requirement to examine a specific type of behaviour owing to the vast quantity of data generated _ _ _ _ — _ -- - - f ’ - I 4 I • ‘ a 1 p — a R k ” P 1 0 it ‘ t ‘ ‘-‘ P— n’- RP cm ‘- — u lp I- i 111 1 11 1 1 Ii 4 i I k P 1 1 1 1 11 77 1 1 1 7 1 1 1 1 I I I O r 1-111 01 —uRPR f l - C f m U -Ym 1 H n b r r nrr D r o I11 m rL3 I I iI 1 i l f j 7’ IlflflDDfl7JDD 1 1 1 1 1 1 1 I-I l 1 1 j 0 O f l ” i l m P O ’ h I I I I Ii’ 7 1 1 1 1 1 1ii 1 1i’ - tY d g b P1l—i m- i 11 1 1 1 1 1 1 1 i i I m r D D m D p b l- I 1 1 I I I I 1 1 1 I I I c I 4-1 ¶ - - -i Q ’-c R - 1 1 1 7 1 I RP - - °- ” - i ‘ I- ’ Phii DT -- 11 I 111 11 P t - k ’ R m ’ Ii III Iii R R 9 R III lii 1 1 1 1 1 V -11 4- Ri’ IIi a- O P 1 • u Ct R - ç 1 — I - — i I I i r • ‘ - n R J P ‘ — - ‘ c - I ii ’i I- ‘ - ii OPk R 3 V VV __ - 7 1 ii ‘ - I - ’- t t C I I h i I I I l H - ’ ’- - m fl - ’°t L P 0 0 i L 4 5 f— T I T m ‘dp r - c m 7 I ’ I 1 1 1 1 1 1 1 1 1 1 1 I 1 I ’ -- k - - P0 ‘ 1 1 T I 1’ t - t L P - f l I 2 1 11 1 1 ¶ ‘ _ 04 ‘ - f 4_ 3- is-- I P ‘ - 4 - I I T I 1 1 1 1 i i i i tt- m ifln - r r’ $ — _ I -0 - ‘ - - P 7 I III I 1 1 1 u ’ f l t O P c • 1 4 • - -m _ -i I I I 3 1 1 I 1 l L P R t • • t 4 ‘ - 04 1 I - ‘ ‘ I 1 b - I- c p - - n 1” fl T 1 1 7 7 1 f - P ° — k R - - I 1 I 7 1 ‘ 71 - I I III 7 R m RL t R i R k I 1 1 1 - l l ’ I ” - - II - -‘ Pk 1 7 f L P P R- 0 l_ P -1 i ‘ P 1 I Z L- r 5 p i - P r-u f- I I ‘ i-’ 1 4 1 •- 1’” - N L ‘ - ‘ j c s- 1- 1 H l u ’ 01 ‘10 I I 1 1 1 7 O P ‘I P - 1 1 ’ P R I T I IT T i l t I t I p 1 1 P 01711 1 - 1 - 0 1 1 1 1 ’ - - ‘ k R 0 1 ¶ kP P 0 J i 1 - 77 7 I 1 ’ 0 IPI-1 ‘ $ I 11 01 1 l I l t - o t i t m D T j f l 1 1 1 1 1 1 1 1’ I T j tm4 ’m u i f l 7 1 7 7 I TI I TI 71 1 III IT I --tt DDI D D’- p L I P P 7 0 1 7 M i 1 1 1 1 l T H ’ ° U f l f l I - - t 0 PI- - 4 l - I’ T I 1 1 1 1 1 P P -J P t PP RP p ‘ T I 1 1111 I 1 1 7 - 1 J I ” r r I 4 lj cl 01 m Li O 1 1 1 1 1 1 I T 1’ r ’ t - I —4 O r - I c I I P P - 0 I I I T I I 1 ’ 7 11 II P 4 PR 1 0 t ” ii ‘ I T J J T T I 7 fT H - 1 it i 0 I I 441 0-” w P l 0 I 1 I i 1 1 1 OH’ f 01- 7 1 0 I 4 711 1 11 1 ‘ 1 ’ T o c rm ruic7 1 n 0 I m -c i p r a 1 1 1 1 1 7 1 1 7 111 j o P P m R — i - I I r L I P fl P 1 1 7 fl J D I I I I I 111 1 1 I ii — ’- l D ’ 0 l-1 4 -l ’i’ R 1 1 1 1 1 1 1 1 1 1 1 I - ‘ q p RR - R 1 1 1 1 TI I I I j 1 flP F i j f m 4 t - j e u j i u i ilTi I RORI f 11 5 11 11 1 1 1 1 QP m J ’ I t P P L 0 k m 0 t C 2 - • -‘ -‘ - - — - I ‘ • 4 I — • 1 - 0’ - - 0 5P P P - 14 1 en P P I f 7 4 IR 4 i I f P n 0 I 1 -_ opt pn • i m- c- m p o’ I r i R c- o I R P R 1 1 1 1 1 1 1 7 1 1 1 1 1 0 4 ’rD ’0 - R 0 R I ’ R 1 1 1 1 1 T 1 1 1 1 1 1 1 1 H 7 u r u p 1 ‘1 1 -J P L i 1 I T I I III I TI f flfl I I 1 - d f ll1 t- I ’ nm - u m P R - -i n op o ’ S P 1’ p 0 p PS p -pm - - ‘t’ h i M 111 1 1 1 1 7 1’ 1 1 1 1 1 II I N I P RI 1 1 1 1 1 1 1 1 ii PR T I T I T i ot t ’ 0 1 I - ff11 ‘ m i P 1 D - t I C D P P ID lo-r l m - p 000 ‘t I 1 1 1 1 1 1 1 1 1 1 f 3 p - n ’r- p p r- I P P 1-0 1 1 1 1 1 1 1 1 I I I HPI iOi DPr ’ 0 0 PP PP Of - PP 44 1 1 1 1 1 1 1 1 1 1 I T 11 IIHD00000fI DO I 1’ ‘ 6 I ‘ R- P- ° - P- ’ P P’1 P P I T I T I I I ’ 71 ‘-Pa flt lI fl I flPD I p pi PI p r I I I I I I M I f 1 1 1 i I r H I f’J0Df JD1Df 5 I ‘ 1 f- — P i- L 0 i O P 1 1 1 1 1 1 1 1 1 1 1 t i c i r 0 D0 I • 1T 4r r f1 ’1Iwl I ‘-s 1 p - i P n - ’ I - p m t- 1 1 1 1 7 1 1 7 I r 4 p - p p S 0 0 f - ’ -4 I T T T I 1 1 1 1 1 1 T I l T I P fl000 V11 PPu T 1 f l 1 0 f rD P I 0 ‘ I -l kt - 0 0 P I I R L P D R I 1 1 1 1 1 1 1 1 1 1 1 P t O PP 1 P R P P R R I I 1 1 1 1 1 1 1 11 11 I I I D 0 L 1 L D - ‘ J c IQ m-’Q -- P R P 1 I I I m T H I 1 00 1 I 1 -‘1 1 41 c o I L t T 7 1-1 I P P I’ 7 P R 4- R • - • — 1’ 19 - P - 1 1’ ta I p -mm am i 1Pm no- I I- H im s-- f T u - I I $ I I I I I I I I I $ I I I -- — 115 — I- p 4 0 p 0 C a — a p a’ — a 5- a a o a 0 UI a’ • I’ • $ I 0 I I V 4 - a a 1 ‘ 0- 1—’ Ia o 41 4 ‘-‘ 4 O 5— 0- 4 a 4 p3 3 ‘-0 — U 4 — U a - V 0 J 0 -1 ‘ - P- • a’ a a’ — s- - p 5 a — Z U II • a I • • i 0 Z — UI • a’ UI UI Z 4 IA a s— - a a I • 0 VI — a’ o - ‘ 1 — — p a - a a 54 5- a a Ui 4 a — a-I 4 01 p 01 a VI - “- a r a 0 a C 0 a 0 0 a a 0 ‘ — a C - ad — • I II 4 5 A a — a a a — 0 a a 0 a j a C4 0 0 — - ‘-- 0 4Z — a- ‘ 4 44 C — I— 5 • 01 1 ‘ —‘I I 4 IL • a 4 U a • - I 1 ‘ U a’ — a _ a’ U I a _2a • • a 44 ‘4’ C a • ft —- — a - U 4 I • 0 4 — 5— ‘4 4444 a 4 • ‘I 0 Ci V 3 a’ 0 a a 1 IA • 0 a 44 U 0 — a s41 •i 0 — — a c a - U a a 4 0 a a • “4 4 _i a a 0 1 • u a a • 0’ • 1 1 - 0 a 5 a 0 a 4 A a • ‘0 In ‘ U a • U I— ‘ U a LI a c • a A 0 a a — a ‘ 4 UI u- 0 a r-a’ -3 UI a’ ‘ - VI U’ — — a t- a’ 4 ‘4 VI -0 4 i ‘ — t j a — ‘ 4 a “ a “— o a- LIz a D 0 a C o 4 4 ø - 41 WI a 144 a • I 4 1 1 4 5- U — a I - ‘4 w 4 U 4— a’ I- u 0 C a • 0 0 5- 4 • I A ‘V w a a ’ a — — — 0 —- ‘ 4 an a • — • ‘-a’ ‘A • a’ ‘Id • a’ • — — a- ‘A - LI 1 0 a a 4 4 4 U - • - - 0 4 a - 3 3 •- — a —z z a 4 0S 0 p - 5 — — 5 01 a — a - - - - 4 p_ 5 — 5 - 0 U - U ‘ — • - LI a a --0 u-i 1 - - a • a — 0 a U t ___— a 4 r 5- ’ a a a p-- LI - — U •- a 3 a ‘ 3 -‘ a -7 ‘--QI 0 4 a 0 U - - UI a W w ‘- -z ‘- a 4 0 0 a ‘ 0 - 4 0 a a a -‘ U I’ P ‘ 4 I a’ a t- 0 a-- t I l • I-4 4 0 Z 0 VI — I- — • a U 4I 5 0 • ‘ — U’ a 0 _ a a C 3 0 P a LI • Z 4 a a a VI a a 4 0 D a — a- Zl l I 4 4 0 — ‘ a a a’ an Z a a — 4 0 - L u a a “I an a — — I#5 a a a a UI 4 Z a’ — — a- U’ U p — -u - P 0 a 4 a ‘S a 1’ - 4 a a — 1 a a 0 j 4 a’ 0 4 a • U a a a 0 a — a — p a I I 0 • p I- — a 5- — a • — I p I U U’ a 0 0 I a’ - — a a 0 ILl 5- Ut a 0 ‘V — U U C U Cj I4 UI a 4 S 4 — a” — 5 - __ — ——-----— — C’ ‘ -- 00 - - I - - — - ‘— 3- H — V - --- V - - MESSM r RF NM DFLAY TLME a - S i - I 4 a 7 9 10 It 2 13 4 15 16 17 71 1 6 7 1 21 1 31 I 138 83 I 1 5 t 7 1 8 1 42 1 7 7 55 1 36 1 87 t ’ 1 46 1 2 6 1 2 ‘ • 21 2- 2’ 7 11 t 8 1 7- ’l 1 7 • 1 8 1 1 19 1 74 31 ‘ ‘ 41 0 80 1 2 0 1 6 0 2 00 $ $ A 4 p $ b $ $ $ $ $ $ $ i t v $ $ I I b b S $ $ ’ $ SI $ 4 $ t - - $ $ $ $ - a ’o ‘ iS ‘ - 4 m s $ 4 s s 4 C -S 4’ i - 1 4 t 1 13 77 1 4 3 1 5 1 85 • •y7 1 5 0 1 43_ •f 1 81 1 8 2 F iq ss $ s s s b 1 — $ $ 1Or I 1 16 t - ’i I 0 40 S A I L T L M S IN S CS A u t 5 f r 0 4 t 1 A # $ A - $ n - S t I5 I I5 $ $ $ $ $ $ 3 11 Mes sagu RFN i Uc l4y ‘I’ flYC ‘ 1 I i o - - _ _ - - _ n Sccs — V - - CHAPTER 9 1 9 V V _ — EXPERIEN CE WI T h UK L C i’ V V A RPANET Introduction Use of ARPANET from the UK during pat tern that was bec oming established 1976 has largel y f o l l o w e d Lh e in 197 5 there is more h o s t- to—host use for collaborative project s L tween UK and tC resuurc groups and less termi n al access for more c asu a l u se T h u s w e were able to reduce the n umber of TIP ports in the PSTN to 4 b u t channels to the RL 360 were frequent l’ - concies ccL The requiremen t for hos t c o nn e ct io n s to- he UCL TIP resu ts f o a problem which has inhibited networ - u se for th a t of n u m b e r of -ì inpu t and ou tp u t w i t h on i y termin il ac ce s s was discussed in the users report 1976A for 1975 Chapter Ot her problems have ar i sen t hi s year which node and the at tachment of the P1 360 t gr- ujo s This problc 7 of K i r s t e in a re p e cu lia r to our the network Becauso- of our sl ow link to the res t of ARPANET -‘o- hive a 9 6K L- - s Lin e tc Norway compared with 50K bps lines over most of the network and t € growing c o n g e s t i o n of t h e t r a n s m is s i o n lin ks tCcr€ vc re t i m e o u t problems w it h the num ero u s Tenex systc- rm w e use Fills - - oan t that It was v e r y d i f f i c u l t f o r us f o r a time to a c t u a l l y estab ’ ish a Lc ji - to a Tenex This problem was solved by thc- extension of the releva n t Te ne x t imeout pe r i o d to allow enough ti ie for a response to I c received from t he UC L no d e Our method of attaching the RL IBM 360 1 5 is a host to the ARPANET has been t o implement all t he ARPANET i rotucols in o u r ioca PDP9 which simulates a remote workstation to t he 360 T hi s h a s worked very well but some disadvantages have becor-c noticeable thi s year One major snag is that w e have to implement our software t erms of t he 3 60 operating s ystem over w h i c h we have no direct in con t rol our ARPANET F ile T ransfer Progr a m has t o submi t t w o jobs to the 360 to get a file transferred US from R u t h e r f o r d to a site This can take a few minutes or more a f r u s t r a t i n g user An other d i f f i c u l t y to Jet wait in the for the is t h a t h a v i n g t o go through the UCL POP9 to the RL 360 means that there is twice the p r o b a b i l i t y o a failed connection this makes remote interactive use from the US less attractive V V -- Finally the operat ing sys t em at RL was n - t de sioneC - -—— - - - - V _ V _—- _ V r for ei cte interactive use arid it h e hoes d i i l ir u i t f o r a n o w L5 U S L t to f i n d h i s way r o u n d the s y st c - J - An in uroved f i l e t r a n s f e r ihe se probiems are all b e i n g tocj l d J r V g r a m w i l l soon he a v a i l a b l e which should at least make the prccess les s frus tra ti ng t o the user by rc-laying messages an the progress of transfer t c Online documentation is bein q put n the 360 which wi i help users get more information about the system and it is ped to improve Help facilities when time effort is well worth while f o r The a i ti es erri d ts sp ite of the above d i ffic— in collabo rative research between groups using the RL 360 and sites in the US is proving very valuable to b ath parties The net- work provides a means of truly working together which was not in any way possible before The p a t t e r n of use u s a whole from the UK is anal ysed in Section 9 2 and c o m p a r i s o n s a r e made w i t h g rou p s use a r e g i v e n i n S e c t i o n 1975 D e t a i l s of 9 3 The information is defective in two respects usa -c individual First the UCL is not mentioned most of the rest o f t h i s report covers that Second the information is based on the utilisation of the t oh -i o retwork by the UK research groups nt on d i r e c t f e e d b a c k f r o m US c -l 1b ordto rs The UK groups a re obli gated to furnish us with r e s ear c h and p r o g r e s s r e p o r t s thus making the writing of this - - chap ter possible It is impractical to obtain the same quality of feedua-o k from the US collaborators directl y I- r a i l y in S e c t i o n 9 4 conclusions are made from experience so far about the usefulness of the UCL node to UK and US research 9 2 A n a l y s i s of U s a g e d u r i n g 1976 This report considers the act ivit -s of 34 research groups for whom appr ova l wa s r e c e i v e d f r o m AR PANET their p r oj e c t f o r t h e i r u se of the Governing Committee of the UCL t h e A RPANET node i n d i v i d u a l r ep o r t s s u b mi t t e d I t is based on t o UCL a t t h e end of 1976 A n s sI-r of other groups were also approved but so late in the year that i t is too soon to m a k e a n y a s s e s s m e n t of o f the t h e i r use 1-1 groups considered 10 may be classified as being ‘h ighiy- — — 119 active ’ 12 have ‘ low—act ivit- ’ status inC the remaining il have made virtually no use of the network during t he ye T h i s c - o n p a x e s with 1975 when only 7 groups out of 30 wore highly—active in their use of t h e n e t w o r k a l t h o u g h o n l y 5 groups were ncn— a ctl ’e There is no predominant reason why so many groups have been inactive this year The reasons vary widely and relate n c-stly the individua l situation some have been held back by lack of or money some by illness A few have L o o n they needed on network mach in e s - u na u c to n e t to tir o ‘-he aco us ua l l y w h e r e tnc-y were u n a b l e t o establish mutually beneficial c-antact s w’ th Ub feiLm w research w r- — ers with ARP A contracts The distinction between high and tow act ivit ci t h e netw -ark is not - al w a y s easy to m a k e I t is not a reflection simply of the of time spent accessing hosts it is more an indtc i tic-s of the L - - r— tance of network access to the research group concerned A r uot-c- r of users find network access useful but cou Ld r o gr e s s w i t h t h e i r research wi thout it these we classify as ‘low—activity ’ ‘ High— act vity ’ groups arc doing work which would eitser n o t b e possibl e at all or would progress much more slowly without the network The resources a v a i l a b l e on t h e ARPA n e t w o r k v a r y tr e m e nd o u sl y There are th erefore m any d iff e re nt w a y s of u s i n g t h e n e t w o r k and i t is interesting sect ion to compa re these t h i s w i l l be done later in this A n umber of points can be made h wever about network use in general It does seem for instance t h a t a new u s e r that it Will take 3—6 months before he can bc- cm o n l y e x c e p t i o n to t h i s h a s been d a t a b a s e a c c e s s useful s h o u l d exF-e- wor k - The where the instran-- t i o n s f o r i n f o r m a t i o n r e t r i e v a l a r e c l e a rl y l a i d- c u t the s”sten m d h a s been d e s i g n e d f o r easy u s e There are a number of reasons why it takes lonqer tm estabilsh a means of usincj a computer on a network than on a local r achine It may take a while to arrange an account on a remote mach ine havmr - established an account there ma y be delays in getting d ocuren t at -an a n d in forma t ion on using t he sy st e m to ARPANET have operating a ctive systems which were n t d c - s i g n e d terminal access and are therefore diff icult to use import a nt of all if Again some mach ines connect the’ pr o cc t is i a pattern of use has t o collaborative one it for inter- - - st he established p a r t i c u a rj - - is v e r y u -i sy -- - - to see t h a t _ — -- — A RPANET SITE -- - V- -- - —— - - - — UK USERS ANL Argonne N a t L a b BBN 12 1 - - O x f o r d N u m e r i c a l A L g o r i t h m s C i-o u p t O xford Nuclea r Physics Bolt Beranek Newman Med ical Res Council Queex ary College CMU-A Carnegie-Mellon U Queen Mary College Royal College of Art L i v e r p o o l T Datacomputer Thames Po ytechnic - HA RV —l 0 H a r v a r d U O x f or d U N u c l e a r P h y s i c s R o y a l C o l l e g e of A r t ILLIAC IV S a l f o r d U R e a d i n g U ILL—NTS Oxford Nuclear Physics Queen Mary C ol l e -3 e ISI Illinois U Information Sciences Blacknest Research E st B runel U Cambr id-o cInst Cornput er Lab Ministry of Defence N Lond o r P o l y t e c h n i c R o y a l C o l l e g e of A r t Readinci U ‘ Salford U LBL Lawrence Berkeley Br i s tol U Durham U MIT-Mul tics Massachusetts Inst of Technology Royal C olle g e of A r t Lab NLM Na t iona l L i b of Medecine Br i t i s h Li b r a r y Rese ar ch D i v i s ion B r i ti s ’ Library Lending Division Dep t of Industry Office— i RAND— UNIX Communication Studies Group Roy al Colic-ge of A r t Rand Comp Queen M a r y C o ll e g e T h am e s P o l y t e c h n i c SU Stanford Li Edinburgh U Computer Science Dept SUMEX-AIM Stanford Medical Centre Royal College of Art UCB U California Berkeley Queen Mary College UCLA Queen Mary U California Los Angeles TABLE 9 1 ARPANET sites used by 1’K _ _ _ _ _ _ College u - I 5 _ _ _ - - - — M ode of 9 2 2 122 — — -— ---—- — ---- — — — — Network Use As m e n t i o n e d e a r l i e r t h e ways in w h i c h UK r e s e a r c h g r o u p s use the ne twork vary quite considerably There are basically four types of use at present Database access use of teleconferenc i ng and message facilities for the exchange of ideas and opinions a ccess to software and or data at a remote site t ransfer of files from a remo te site to a local one Some groups use only one method others may use a combination of methods Da t abase access is made by t he g roup of u ser s who a r e p ar t ic i pa ti ng in the British Library CANCERLINE Pro ject and by the Thame s Poly- technic Project CANCERLINE is available via the NLM machine whic i is not connected as a proper hos t to the network and is d i f f i c u l t to access This difficulty has been overcome by a program written at UCL which makes the connection for the user CANCERLINE itself is a good s y s t e m and has been used s u c c e s s f u l l y w i t h o u t any f u r t h e r problems The Datacomputer a mass s torage dev i ce is st ill l a r gely an exper imental system and requ i res cons idera b le time to become fam iliar with Access to it is therefore not typical of a database system the difficulties involved are similar to those of other newly- d eveloped s p e c i a l resou rces such as t he ILLI AC IV pa rallel process- or Resources of this kind are so unique and so new that the intrep- id user must tread slowly and carefully around the system until he identifies the hazards Use of the network for teleconferencing a n d t he exchange of ideas is now so well established that there are few problems involved Memb ers of the UK M inistry of Defe nce have collaborated w ith the US Department of Defense in the evaluation of the MoD real-time language CORAL over the network Wilkinson of the Medical Research Council Ca m b r i d g e is collabora ting with Donchin of Urbana on the analysis of rieurophysiological data using teleconfere ncing methods There have been no problems excep t t hat teleconferenci ng can be somewhat tedious using a slow t erminal The mos t satisfactory way of using the network from a terminal to access remote software and data This u s u a l l y requires only a is small amoun t of input and if the program being run involves lengthy computations any delay in response from the ne t will not matter too Brune l U and Cambridge U Computer Laboratory have used this - — - - - much - - - - — —- - --- w - - - - --- - -- - - -- —- - —- - - - - --- --- — 123 — method successfully If the output fro the remote rr -qram prccia2 s large files there is the problem of getting it into hard t m o p y f e 1 without tying up the terminal f o r h o u r s th is pr o b l e m h a s been l ar g e b overcome f o r UK users by t h e a t t a c h m e n t of a l i n e - p r i n t e r to t h e TIP and using this to print user files The fourth way of working is to transfer software or data fr o - remo te host machine to a b c -al one and use it there for develoto- c -ot and testing This method i s used for the high-energy p h y s i c s collabora tion between Oxford Harvard lilinois and Chicago- as they have encountered some frustration in using the RL 360 195 from t e US The p ro b lem stems from having to rel y on tlii’ce computers all b e i n g o p e r a t i o n a l at the same time i Il l no i s UCL RL With wide l-c differing local times a t each end d i f f e r e n t m a i n t e n a n c e p e r i o d s together with broken or aborted connections it can be difficult t times to achieve a useful work session T h i s situation has been repor ted by only one group howeve r and should no t be ta ken as typical network situation o If the only v i a b l e way to use the netw -o-ro was to transfer files from one site to another then it would be difficul t to justify its existence in terms of resource-sharing Flexibili ty would appear to be the keyword ploi ta tion of the network ’ s facilities Most projects develop by sta g es and each stage may need a d i f f e r e n t type of resource It is interesting f o r s ucc essful ex- approach oi a differe nt to see how two of our user grou p s Br istol U and Durham U have used the network this year Bristol began by transferring programs written originally on the CDC 7 600 a t LBL onto t h e RL 360 195 where fu rt her developmen t too’ p lace They are now entering a second stage of work For techn icml reaso n s the next stage in the program is bes t ru n a t L BL whe re the facili ties are better for their purpose They are currently devel- op i ng the programs to out p ut the a na lys is f ro m the 360 i n to a fo m f o r t r a n s f e r r i n g to LBL f o r f u r t h e r a n a l y s i s At t h e same t im e Kelly of LBL scans the resul t s of the 360 runs from LBL and keeps a log of the status of the runs W ork i ng in a similar way Durh am U have implemented a version of the LBL Da t a Management Program BKYBDMS on the RL machine i mplementation it was invaluable to be ab l e Dun n-i to test the machine in d e pendence of thei r I BM subroutine s by r un n ing them on the LBL C DC machine - - A second stage of the work has involved -- - - - - - - ‘ - - —---- - the desi gn of -- - - - -- Pu —’ y - - - —- — -- ---—- —— 12- 1 — a c d ing i i n j u a y c — - -- — ---------- - - -— - —--- — - - - --- - - - — ‘lli lo nas been set up at LI3L to enable others t h e c l l ab o r at i ve g r o u p to add data and carry o u t e x p e r i m e n t s F u t u r e use of the net f o r D u r h a m U w i l l i n v o l v e the q u e r y i n g of t h e in da tabases at LBL for book—keeping purposes Ano t her group whose use has d evel op ed int erest i ng ly t his year is the Blacknest Research Centre way They began in quite a modest by accessing seismic data files at ISI Now they have set up the ir own se ismi c database on the RL 360 19 5 a n d re gu l ar l y e x c h a n g e da ta with seismic centres in the US For them file transfer is the i de a l way of w o r k i n g since t he files a re sm all an d t he da ta is needed l o c a l l y A fina l point needs to be made about the use of such as the ILLIAC IV parallel processor special resources This is’ such a n expen- sive resource that compilation and debugging of software to be run on it must be c arried out on an o t he r hos t machine a n d t he n t ransferre d to the ILLIAC for execution This type of work can be c a r r i e d out f r o m a ter mi n a l on a s m a l l scale bu t any large-scale development of programs r e a l l y needs to be done via a l o c a l h o s t so that li-stings and dumps can be o b t a i n e d f o r e x a m i n a t i o n T h i s of c o u r s e w i l l tend to apply to software development on a n y remote machine but is espec ia l l y tr ue of a spec ial i sed reso urce whe re so many probl ems ar e likely to occur before the user comes to gri ps with it In summary full use of ARPANET is best aenioved tiarouqh a local hos t rather than by terminal access A flexible a p p r o a c h to the way in w h i c h on e wor k s w i l l e n a b l e the reso u r ces and facilities the ne twork h as to o f f e r to be e x p l o ited mos t f u l l y 9 2 3 Problems Encountered in the Use of ARP ANET 4ost of the problems encountered by UK users mentioned a IV C already been Four groups had some troub ’e in Jccics sing and making use of two hosts the NLM machine and the ILL IAC TV found the operating system of the RL 360 d i f f i c u l t US users have also to use The lesson to be learned from these experiences must surely be that when the decision is taken to connect a machine to a network adequate consid- eration must be given to providing Help fa c l i l t iL- s and suitable user interfaces to the system - - - - - - - - - - - - -- - ___i _ _ _ -- -- - - — - -- --- — -- - - - --- - - - - — ____________________________________ - t h e la c o Use of a guest account can Several groups found their work severely an accoun t on a su it a b le US machine h a m p e r -ed L ’ usually be arranged for short term experimental work for serious use a UK user needs to establish collabora tive links with f e l l o w research workers in the U S - the network preferably before he contemplates usin- The i n a b i l i t y to cope w i t h la r ge I O f i l e s f r o m a t e r m i n a l i pc cc full collaborat ion for two g roup s Queen M a r y College and Cambridge U Computer Laboratory This problem should be solved when EPSS i-s in full operation see Chapter 6 The only other problem of note was the slowness of response fro the net This is not so apparent earl y in the UK day but becomes a nu isance by early afternoon when US workers start to use the mach ines On the w h o l e most u s e r s f o u n d t h e n e t w o r k r e l i a b l e a n d most of t he hosts easy to use 9 2 4 Mode of Access to the Network Out of the 20 users activel y using the net this year 15 from terminals 4 via the Rutherford leased line connection to the L a b o r a t o r y ana di so 1 g roup has a Of the 15 terminal users groups admitted to finding this type of access unsatisfactory he ause it LCL TIP was slow and inhibited input and output Even for fast terminals I O i s restr i cte d by very i l m it ed buffer s p ace in the TIP The o n l y typ e of ARP ANET user for wh i ch term i nals see m perfec tl — adequate is for database access or for runnin g programs when I O is minimal There is also a problem of poor quali t y t ransference of da t a over long distances via the public switched telephone network distort I O to a degree where work becomes very difficult 9 2 5 This may Connection Times The M e d line grou p used most connect io n ti me throug h the VCL TIP since there arc about 15 participating groups and they carried out searches on a da ily basis Apart from this mos t use rs were able to carry out their work quickly The group reporting the longest average weekly connection time was Salford U with 6 hours a week — On c L other h a n d I S tel U used o i l stuL eS of t heir pro ject 30 i iru te steraqe a weuh th - c ur l y f L- -imtir-c ti -sii - — 1 26 t ins- mnd eo-ec U i r j h ly a ct Uccau se i n i t - s ü c-in nt- - t ioti the of 1 -atterts j c- t ii ii u o of t h e n e t work t h e e x i s t en c e c o i l a s o r a t iv e l y w i t c - t h uWo rs li- i t -c t o h e d g ou s us i rvi t t he IV w l - ige of ju st beginni ng serious won k t h e - e - p c iL i n c I I t a I U se i m i t’ l i p r t been v e u ma de of 17 u sing k at c i f us - v It i— u a S ’ t suir h i s s g - I - U o - lIds S -i us i i i g ti c-I fl - F e u I T i h i t L a w r o c o P - e r - It ’ hr a oi w h i c h 1W misr S vi t h e r u se i ut e x r - T - f j c - n t o t h i s -ar c t h e n a t - c sy-ent a i - - n - i t Oc g-s -tt na th is - 13 t i n t o l es t i r ye -lo l It 15 I S greats have in e r t i e J Inc Th e n - t in I LL I AC as such aL- o - l utx - l y 1S Ofc Ut D h a V e U- con e - ‘n u -o- i u - - r t a f l t iboi i tui icc - 1 - i t 1 0 - Wt f a rid Ar on - e 1n 1 ort n t ccl til - Th e use -ru is around 2 ii o- i s vu ii S imm ar ’ Use of the c-t- - -rk this ye-a r tUL- tut os a - ° c i- nrc be s -ni in q no mo -e t h i their collat - -r i -i e-C Q 2 ’- f to t s’ t r e i J n ex t icr - - r— --y - od u u - - a tna ctiirR - 1 t l A r 1 Server icJ lit m o s t e c- i i and 5 are in 5 t and iii csage sy st e r c ce he t tee -di t i c u l t u 3 L- ih i s S O C t it - i d r - -ups wtio thei r in -f f le v th t i - r e - i i t i d - n ’ - the ni -c i s Y r e s e ar c h lm- t v e use-i the netw rk to aid l - a - - m ita j ie j dt-d j u t — h ig h and I— a ’ a c t i t ’ i f - c - in -as x t - - n t of t i c - ir u n - o f t h e n e t v - - - k A t n A J U ’r woil—e - st tllislo-d r t - i wh i aurilo l h 7 i- v- hole- th e l m i q h — a c L i v i t On thc h i v tJt - to uSe t h u a et - d - - -t it t n ir e tn - I ti on ‘i n s i s t s th y tccta t u t u th e n - u t w c t 1 n- h-c I of - toui s over sit th-y he c-xpc i rm— to -ed e Til- icki ec t a nd the I esc ir c h c irliui i Ii fl i Veu i t 105 1 - t m r ictul i i l t o r d i u d rue int o thi c - t te qo ry O x f -- t-J U n i v r s i t y i -l t r hh -s i c s r o u y I h e i r c o l l i h e r i t l O f l W i t h C h j u n u o II m r v u r d a n d 4 V e also ‘e n t i r t a - c I i t 1 mcu i l us the way s i i i ri have h a - i a j - r 4-is terms ire i r J - c’t- s u s i a - i A Rl’ ’- N i i Resu- r ch tl in uri t I l l i n o i n be gun in 1 7 t h e f l u i t ish L i I i t a r - ’ Research -ri evetc i- d _ meri t u i v i s ion l ive ce t i r iu - - I t e l r cx - i nt ent ti i n t o r n i t i o n retriev ii pro cut usi ng th e CANt J- RLrN I cyst em t h i s y - ’ar inst - -si ot h k i d 1N1 — — - - —-—- —-- — - --- —- - —— - - - - - - - - - - - — --- - — F -- — -- - — 127 — - - Two o t her groups Culhaiu L inera tory a rid tne R o y a l S i g n a l s arc R a d a r E s t a bl ishment have been -a at vc in another sense T he y have beon work ing hard to establish their own mach ines as hosts on tu e ne twork for the purpose of collaborating with specific g r o a n s in t h e US 9 3 1 these w i l l be d e s c r i b e d f i r s t High--Activity Projects al Signals and Radar Establishnient RSRE The RSRE at Malvern ii s a GLC 4080 computer which runr the reatim e language CORAL The latter is being evaluated by the US Dejart- ment of Defense as part of their yrojcct t- examine various realtime languages in t irms of m i l i t a r y requirements In order to f a c i l i t a t e t h i s e v a l u a t i o n t h e RSRE machine has been c o n n e c t e d as a host to the ARPA network via one of cue -- UCL PDP9s see Section 2 5 This c o n n e c t i o n was p a r t i c ul ar l y u s e f u l d u r i n g March when a g r o u p f rc n RSRE visited the US DoD to give an instruction course on CORAL the use has c o n t i n u e d however d u r i n g t h e r e s t of the y e a r Cuiham Laboratory The Cuihant fusion l a b o r a t o r y i is e m t ories in the same field and woti iis 1 - ta - u elop software to a l l o w t h e C u i l m im I C L a host to ARPANET in order Section 2 7 to i n t e r e s t s w i t h US l a b o rat - -jo iri g on this year to dev- s y s t em facilitat e 1 to be connec ted as 4 72 -S collaboration see Black n es t Research Es tabl i sh ment The Blacknes t Research Esta b lis hmen t ’s project to exchange seism ic data with centres in the US v i a ARPANET began in quite a modest way in 1973 when t h e ARPANET l i n k was f i r s t e s t a b l i s h e d a t UCL They began by accessing data from the Seismic Data Array Center SDAC W a s h i n g t o n and t h e n began to set up t h e i r own d a t a b a s e on t h e RL 3h 0 ’ 195 f o r access f o r US c o u n t e r p a r t s I3 lalncy 1t476 now being accessed by the Vela S e i sm o l - i c a l t h e National Earthquake Information Service This data is Center SDAC and soon hUIS Th e la t ter is t he foremos t collec t or ot ear t quake data from diverse sources Computers and computer netw orks are opening many oppor tunities for improving the quality of da ta which NEIS collects - -- - -- -- - — --- -- - - -- -- - --- - --- - -- - — - —---- -- — by participating in the an d — 12 8 r l tnsn iission of dat v-r i- 131tcknest is assistincj in a project th at w ili nocd tion of import ant ly LEIS have a Lways By tr ansnt tt quickl y than diff erent tm g rnii t Ut f o r e 5C but rutu r i r u i r c c t— B lack ne s t c ht a i n c - far a cr e - 360 are in a f i l e s or t he - t h a t u s e d by iLlS so 1 3 L i ’ k i t s t fron suf t-4 tre wi i c l i from The c u r r e n t data before f urnis t d at a data vi a ARPAN I-ti it may be itig lie --a r— re -ui time it i i rt ii -j uik e s r e - c e ived netaa r ks ass outer s1 w il ud log i n s t i t u t e C a l i f o r n i a ar e - trn lemen— i n t i e s s t a n d a rd L E I S I re fe m at Liii - d a t a 1’ l’— 1 0 at I t tc t i e I SI t in - i i fo m a t ion Sc iences For their part Bt a ehri est have I-t en re ’c l y i n g NO i-WAR bulletins fr-sri the Norwegian Seismic array Unfortun ately oroUuced at the end of September due t o cuts in t hese ste g-cd being t h e i i- s i r B l a c k n est a r e a l s o now reccivin ’i dali’ b u l l o t i n s budget station in Ottawa u r e ly is f u l l y q u i cs l y e ccu i vo R 2 A - L -i i r u n I v ii r i’ i t l v u -i t h R P A N L T the y i n t eg r a t e d S C i S r u j i ’ ej i i c e n t re tie fr-or a These a r e r e c e i v e d v i a t h e V - La C e n te r In due coursc- make the d a t a a v a i l a b l e i n f i l e s or-i t h e I SI mo ’hir c w h e n i L lS r e s e ar c h 1 B l a c k n t - c t ci a i n dat i as 1 0 5 t ha t t ei r t h ey c ur r - t— a t t a c h m e n t t- re - ce - i r e tnt - l i aid s c - i t i_ On 1 he e to get much lur id L I L n l v e ’ I l c oc k of t h e 1’h y sit S S e - j a i l at- nt U r i s t - i h - i s i - t el us i - A R P A L U T t he i w r u c c e B e r k e l e y ccli i i - r at - - w i fli K - l y at c i i the RL 3 6 0 1 9 5 t Th e y ar e en g ag e d n a joint project to d- u u r t o rt - C a l i f o r n i a L I I L an a l y s e p i o n — p i o n to nhie ARP ANE T h a s e na bl e d the CDC 7600 taken Place t- n — i n t inu c l e an s ca t t e r I n -n d a t a t h e m t o e s t a b l i s h p r o g r a m s w r it t e n U s e of the o r i t in l l i t LBL u n t o the 360 where f u r t h e r development has Total connect time for this was only a few minutes per wc c- k on average I l ey a rc flew e n t r e as o ns t h e n e x t facilities output ti e- eri lig t ep -i re l el I i- r in i ly sis in ti n- I h oy at - - - Tb is p a r t without t h e network -nid the ii iau risely useful tar p i e gr a t u i fro p tin ’ 3 for further analysis F’- i- tech n i c a l eeand stag - - of work a k - l I t - il 1 t e s t run at LBL where the curi - - -nt ly develop in g i n to of 1 ity sca n -i for fir th e not t r a n s ft -i- files r e s u l t s of pr ogram t- t r an s ferr in g th e - work c -u ld t i the t he t o LEL is contemp lated It is also 360 runs from — - - — i ” - — LBL and to keep up—to—date log sheets a the status of the runs British Library Research and Development Division The experimental project to access databases and the information retrieval system at the US National Library of Medicine has c u n t i_ n — ued this year w ith the modification that t h e database used has seen CANCERLINE instead of MEDLINE Comparisoas of the two databases have been made by the user centres participating in the expe r iment These user centres must record detai ls of their searches for BL records in return for terminal access to the information retrieval system rheir pattern of usage of the database has been m e a s u r e d also by the UCL group to corroborate the data furnished by the user centres These measurements have been discussed in Chapter 8 Cambridge University Computer Laboratory Fitch has been engaged since 1973 in a c oL l a b o ra t i v e p r o j e c t o v e r ARPANET w i t h H e a r n of U t a h U T h i s work h a s been c o n c e r n e d w i t h the comparison of algebra systems CA’LA L an d tively at Cambridge U and Utah-i U REDUCE prodaced respec- In let-eiident comparisons of each other ’s systems have led to substantial improvements in algorithms and t h e d e v e l o p m e n t of new s o l u t i o n t e c h n i q u e s T h i s p r o j e c t was essentially completed by the end of 1975 Work h-a s begun on a new project Fitch has taken a copy of the scc -n -i version of the LISP compiler over the n e t and is currently u sing it t ie h This cooperation does not require much machine time on ARPANET but relies heavily on file transfer facilities ued It that EPSS will make two—way file transfer much easier They h ive continued to use MACSYMA u s res earch Ott a i is h e l p is their algebra number of occasions they have been able to use tu in - i jue resource to considerabl y shorten the development time of re- - algorithms Jackson Of particular note here is the help they afforded J a c kso n 19 7 6 A packa ge for Petrov classification was implenented on MACSYMA and as a result file space was made available to their group on MIT—MC r 130 — Surhan — Univ e r s i t y th e - Phys i - s Depar tu nt D u r h an U heads the UK section the Intern - it ion -a l Pa rticle i t i Croon sis wor k is ccncern- j C ta 1 t -at with the -Ievelopm ent of c ner L iy h iuh p a r t i c l e— s c a t a e r n e l e m en t a ry d it ah a s e s t he they have - ‘ n t inu t ’d thei c collaboration east y- - - in a tb burm the i a ri -flee Berke l c-y Partic I is Data Croup at -urp-ase of the w o r k l-i is hei r in rkeiev Data iani - io i- nt T h i s a s p e c t of the 1- v ’ in irip la- - ented in -r hot sb B KY FD ’lL is u-p l c s i i m p l em e n t a t i o n L i i r oip u - ds indc- uend or c-n - o f m a c h i ne the Berkeley fran D u r h a m e n c o d i ng il-it coop ire most ‘s-nve n ient daLi in - I -ni er mn - ‘s- r the to -mr- - - 1 -a’ - r the tr-an t0-r k- a - -iv- au i1 il ‘ -- --it r e j i I 1 i i f l5 rn i- ’htne h- - - ii in e s at berke loc he aPi ii tv to i-sri ILCU- ‘ hem on i— - nni sc ri -Lum it its hr -ug h ii - t ’ Pr 1 111 Lmn - -nuaq c i a C at The Cal Tee an d - - i r k c a s i 1 i - r i d i r t - s t 1’ in As C a l t e v b y sot t in - - a t h e m the o t h e r s il low in - r in-i LII I h-li- t n I ar fl1 si r - i r i - i so of Ti e- -i -r v i I s Iii i A ii l - I t I A N - i I i ’ nt - ii it up cr - c u p the been save an the t o a dd duti i l1 l - t a — -c t -ni vi i I I U 1O f i t i- ii - I ‘ —- t— t e u r rnln i of tin - I will t i - in s fer - l I O I -- t i a t is n t AR PANI I - -‘ r- t - r i 1 u r u -iH PAi1 I’ ar i f b r - tsar j a r st e tc ISt i t an mis m -j k ha p 1 i-s - - i tb-n design of w i b hi ti n -rn - nt 01 da b - - - t -l L i se s i sur’r Part i el-c Phu s ics Di PPnL n i t -1’ -t hive 11 ’ IiOi - work has involved is Ak PAiI - T iceesti i -- is i - n on a t’D m a c h i n e and dun n-a e -ii nab Li to h e able to t- a t n-c was it iv a l u ab le l a n g ua g e ta r i t - ‘ t Al o niu ’h n o CD C jesi n is le - in g - ‘ar jo-I ht i ’h i I Lou Another aspect of t in 0 1 a first vc-rsion - 1 i r - d i 10 7 1 n d i - 1 - i o’ ’e-d yers ton a - i s An The to imp lenent an IB -1 versi -m n of a fl -fl’BDMS or t so RL 3t 1 9 5 Pro st in t ad in Jul y 1 0 aus running the - ‘ - i t h - ar t - be - the L a b - - r u t o ry - cf ccc rather ft in nt i r k - i 1 i - c exch m n ed by - iii 1’ -- hat i y i n-c 1 0 “- t n he - i a w r -- - - -- - 131 - that without it the nature a their proj cot would be alter oh co s ieierably Oxford University Quirk of the Department of Nuclear P a ics Oxford l e a d s a yroup who are working in collaboration w i t h groups at Harvard Chicago and Illinois This work concerns exp er ini- nt 3 S at the Fermi National Accelerator Laboratory I l li n o i S h u o n— s c a t t e r in - - data gathered at Fermi is reduced and anal ysed usin-c software implemented on the RL 360 195 The collaboration has been working success ully for severa l years Anderson 1976 The main difficulty is that use of the RL 360 f r o m a host s i t e in the US depends on t h r e e m p u t e r s th e CS m i c h 5 n the UC L P DP9 and the RL 360 being operation-i l at widel ’ different local times and maintenance periods This t-anether with Iro ken or a b o r t e d n e t w o r k c o n n e c t i o n s h a s c a u s e - C t h t eroup s to use the n e t c for swapping data and programs rather ti -ian utt eripting t- rk remo tel y control the data processing or detiueqinq of pn- qran- s -c- -h a r a c a aistics of the UCL node described earlier in Section 9 1 have t us affected t h e way in which the network has been used by t e- se c o l L i -- orat ing groups Local difficulties at Harvard and Chicago have also caused some problems there has often not been a suit ut le p rinter available at the time of transfer of line printer -utput Nevertheless work is proceeding satisfactorily files ti-cl close c c n t i c is kept between members of the UCL research g r o u p and I l l i n - i s smooth out any problems that occur with transferrin g fil es Reading University Ilockney ’ s group in the Computer Science Dept Reading C is engaged in a project to evaluate and develop rapid elliptical sci’ ers and particle mesh algorithms for parallel architecture computers The p ro ject involves the s tudy of different parallel architectures in addition to the numerical algorithm s P — - — -- —- —- -- — - — The-v save established an d t v P - t p e - a n assoc iat i-on with L- may Field tm file in C mnsfc- r ti e -i lq- r i t hp i in term s of La - c-aus tin- c L-en- t n ark t e st w a s only as expc-c t ad Tire Latter i is used for time and non— ‘ z u rl ip m o le p ri ct - e j m e a s u re d contin u es ‘or n -u t e r to lie is t h i-ic run is T n s r -i iOi ’i 5 S i i his rah r on than i t i c i- t on hef- ee-n inst m a ted run rest ic ions me - seC ¾- v of t r i s u se of ILLIAC F t le- -iu - t ’ i rr -ti h i ’ t - - ts-s lan qs cc- bef -orc i t s n e r f a r o a n -ce C i li e - can oe ar e - r I 0 ir p i l a ti or i of 360 LbS th a t ar - -i IV h ive been disappointing i v a i l at - l e twi ce If’ for t h e y h a v e an a c e - -u n - a Fortran—bu sed parallel p --n The een a n t e - s i - ’ Jesshope 1976 Preliminar’ results using the ILLIAC times Stev en - s at the I L L I A C t ’ for t h i s return ILL IAC the IL1 AC IV t- ir In time sp ree on Ames ih t 3i0 TSS system checkout of the CFD code s -u i J r n i s s t a n in-i A eel Lii- -na tive proje ct has u s e in vortex c a l c u l a t i o n s n u n - h er — ft32 - n-s i ffet - te- ject since no oLo n a s r a l l o l pr- It Silford I n i v e a s i t ’ C it c m it i c s f i - l is S ti ford it jjlt -— dmno n ior t i ml i 1’ -- i ft - I Ira I - ti khc n a 1 - m n n i t Pm m fir this u t - il ‘P1 in to imi l ‘- ri’ p led I -a ‘s--n - il a - t a- I — e n iI I —‘C I e ar kee n P - tI - 75 Ti - m t k- I f i - r r c -r ni -1 - ’ u - - f u l iii ’s AI ’ P i - ’mi ni - t h e - itt t tn fran - i t im -’ -i l ict in st - -n s to hi cxi-’ r m - n ° - - d y i n s -r i l - i - i - ’ - -jii t - ti m e l i f t r - n it ies w i t h ill 1 1 j a n 1 1 I- tvi l y — l i by is i dcd run t h e user the i o IV i 14 — T e n c x faciLity r - i n sfen e h I-et ’o - -n t o t itt ’ in ’ i I -d n t La P h ire t h e - i - c o n - a n 1 - in n - for 1 ni t faa - i i i - ’-- ’ I t’ r a t - s l i d li J f C fi l e had to lie I lox I i i ’ 5 hi- li l t ion of this car -- - inC di -mnirgod - sO - — 1f m t x in ha- - n - v t -lop ‘ 1 1 7 1- I i ‘k of Iro steady -Ic- i rt - c-nt of ftc- ui-a’ ‘i r - - ra rt lc a l re-t son ’1 5 t e iil ma ’ O n l y oPen I h i - y rr - r o islv 3 -0 j t - r i n ’ i nt j r n u t - red t o t H - f - n - o iii ’s I-ru -try of i r i v o l a ’c - cc-i-a’ ‘- i l cul i t inns that i n - Progr am - s nra’ - 1 t t - n i thieb r n - mn — r - r - - c -s s m ach i n - s t -s e a he IS - r a- h i i ’t ng - fliama supe rsonic v o n i f ’ ’ t i m - - r i t n far- ilt- ’- ’ -I -- r- and i’ p roject file at f ete h i s inC this has t ra n s f e r s have l e n a t h of v that L- t in c s l ow t a ke n t ic r l i n k The work c a r r i e d o u t so f a r a t 1 4 —T e n e x s c a l e t e s t s of allocation of has consisted of small- s imple c o m p u t a t i o r - a l p r o c e d u r e s The f i l e sp ac e 100 Tenex pages has been i n s u f f i c i e n t to alL io more than one job to be processed a t any one tim e si nce t h e ILL IAC sys tem produces large outpu t f i les du ring p rocess ing and hal ts a j o b if i n su f f i c i e n t d i s k space is a v a i l a b l e I t is expected t h a t more file space will be made available soon to ease this situation Progre ss has been made towards a c h i e v i n g their short term aim of obtaining run times for the calculation of solutions of partial differential equations associated with fluid flaw field problems The experience gained throug h performing rather simple tests has also suggested prograrmiing strategies for more complicated problem s Walkden 1975 The next stage of the work is to tcst a full supersonic flow calculation program followed by full scale tests of algorithms ari d c o m p u t e r s t r a t e g i e s i n v o l v i n g t h e c o m p l e t e 14— T c -n e x s y s t e m This w i l l begin in 1977 9 3 2 Low—Activity Projects British Library Lending Division Harley of the British Library Lending Division leading a project to investigate the feasibility BLLD has been of an i n t e r — l i b r a ry loan network between the British Library and t h e Nati onal Library of ‘1edicine NLM The purpose of the experiments has been to develop and test a system DOCLII-JE which minimises input and reporting costs and uses a central computer an d address file maintenance DOCLINE is now operating NLM for routing reformatting s a t is f a c t o r i l y an d a l l the 16 regional medical libraries in the USA are now able to channel their photocopy requests to the NLM computer Requests which cannot be ser- v i ced by NLM are t r a n s f e r r e d to the BLLD f i le from which they are retrieved d a i l y b y BLLD staff 7931 requests were retrieved 1 2 mon ths e n d i n g J u n e 1 97 6 and 5 66 8 in the ms w ere sup p l ied Reports - - 134 — of s u p pl j eL l i t em ’ — mmmd n e a s an s ad Iron the BLLD t ecmn ir m al — for non of ot ers ar-a e n t e r— sun -ni dail y and use of AL PA td- T wa s i n’ ’ti - u aL- Ic rP h m- ut ib t i m e s’ u t e rn works well it throuci h ARPANET has Poet i i i o e t t i n - it esti I toh ed access t i By m u t u - i l terminated TYMNET s i n c e i t Im t s f- t - - onmc no - on cr at e c i v i a vstd i s tin ’ i Tr eemac nt rout inc opera t ion i I3rune l l n v a-rsity Social i’ i hi t w it h six s -s mr es t a h l i s h id w a r e and the a s s e ss i ex mu - nmem - 01 fh e it s ‘ h a s P r u m - ’l ‘ -a mis - I n- it - - a ’ic- Liv’ lie-en n -v s t en a t t i - li r i - s i-a-I ‘- H h I 1 ’ i n - f - - n - i t I n Dep in n- -n t i na t a- m c- c- -m ay li k re v i - it - i i ‘J it t O- t r i t - a ’ ii i- i t i mi - i I h in - I ar an - 1 7 c o n s i i cnal I it 1 m nes OXi - lo di I f i c ul t v has r is e - _ m s inpu t e r r o r s d r mr i n tj i - u- ri h its r i - l i i i i Ilit ’ a n d wtr - - i’s - t - tin-I u n d er --r t s uHf in - Ri - I - ar t - ‘a- i-u ’ - ‘ nO se ’ Apar t su c c e s s f ul ‘-¾ of t h e a nd TenL -n cr t- - and -lit - - i nv e s t i e - a t e ch ’ve -p-r e nts li-iat- tn lii - - r PA - Ll H- - si- ’ ’- -n A I A I1 h i s l i -On oh i i ii C u n m p a r i n -o n h a v e t — -i n RA I tA il on - - - If c i- -u m e - J as lc’ i m ID I OL - n - i m u h ‘irn inotO P t i - c t o ce n the t cr a t a’ i P I 1 N 1 access some ii- Df I NI- sa n-he-is art ’ Apji i rc- n t -LI l A - t i n - e s f u l l -- h i s b een acu ’essina C e n t re Pr -s i n- mn-- i- m i t lIa t in s t i i - I - i - i n t - ek t -s 1- I mis ’ ir - i hr it t i n ‘j i - the 1 tin - Teu - ii r ial iu mv sy t fl - h ose ¾-- as i n d i v i d u a l ‘ s a c c ur a c y net-s ir k mu - tsr -j o e l the - — -f Tn d u s t r P i k n-r t o n - Ic-v e a ’- nm- cter s-a- It— of if been m v l m a T dI —n- i i r t r a e n t I 1 c e - - u n ts sh a r i n L an i ii c i 1 t c - l u - o h o r u tr m ii i i an ‘ l u - - - ’- -d lines I tit 5- l sr u n c l ‘ s p an t -t e r n i t p- I c-s ‘ -I to th e pi - - it rcsee I isa 11 H is p - j - 0 in -u - the of l I e I s e ’ c u r i t e n-u or- -i - C i l I ficili t i ta - - t o -develop C A L I U A been en in ir d e r in a d e c i s i o n a n a l y s i s ISI it O rq n j n -j t i o n an-i at time UK and UtP I riot-i i ’ ’ u u u t s i - In s t it ut e r t i e i r tt Lna is -nc r r ou -s -it tiit - i t Unit ii It r un e l at sm -hi es _ i - oj - - t la ’s An - i S i - u i - i I t ‘i l i c i m - ll sit found -- --- t ‘rn-rn- or to i- c ri-ore sat 1sf mc— ‘- --‘ - - - -- ---- - -- -- - -- -- - - ------- IVI — - - - - - _ _ _ — to r y and in t h e second half of l97 a Lm o s t c a n t i n u o u s l y f o r R AD A B c- r cm s Pal t-2rm ’lroi l h a s Leen ascii I- - ¾ar burg h University A new prnu f — g e n t - r a t ma n i m p l e m e n t e d -at E d i n l u a r - h in system is implemented vers ion in c o n j u n c t i o n tsr tIc Co rd - with rcr f - ’r A nc -w h a s b e en b c e -n designed and has - J ek ir t nc-nt HCI-P IS l’ in t-jined at Stanford and Medical sy - t -I v er s i o n t h s in-f i r iv e s i t j - a t i n g the i-kiinburg ii Science L - rn - p u t e r The i s main— the Stanford i m p l e m e n ta t i o n Research Council It has been proposed by D o n c h i n of be made of the f e a s i b i l i t y toe ‘3 of Illinois that a test of m a k i n His l a b o r i t o r ’ equ ipment av a il- a b l e f o r u se b y remo te l y located im ’estiqator s “ia the facilities of ARPANET W i l k i n s o n of t h e A p p l i e d Psy cho iu -cy Unit MRC Cam br ade e i s c o l l a b o r a t i n g w i t h him on t h i s b a s i s dse n - a f - ar of the netw -arh h a s been restricted by lack ‘-if am er m ’ i n a l a t C a m br i d ge and has bee n c o n f i n e d to the ex c h a n g e of c o m m e n t s on g r a p h s of s ta t I s t i c a l anal- y s i s of d a t a The overall aim is f o r w i l k i n s o n to view and a n a ly s e ie e n v i s a g e s a series of studies in the gener a l area of t he rela t ionsh i p b e t w e e n e v e n t r e l a t e d p o t e n t i a l s d a t a over t h e n e t w o r k in the EEG and performance Ministr y of Defence MOD Curry of the MoD has been using the network fm r collaboration with the US Department of Defense on the H i c h Order Language project currently in progress in coordinating The network f a c i l i t i e s this project and i n providing have proved u s e f u l access from the US to the M 0D language CORAL N or th London Polytechn ic Survey analysis packages are the subject of an investigation Rowe of the London Pn l y t e c h n i c Computer Un i t to t h e p r o j e c t all ot ting him only t en weeks - -_ -- —-- - ---- lie as signed a student fo r the task even - by _ _ - hf — t h ou s i he h i d fa r - ‘l’iii— n - ct tO O 51 u - i i t urk -- - Nun ’ L v i ‘i a-i-mr o’ I- i n - v is ‘i n - f u I no Ly e f lu uiqor u lla i orat l i a r D m v i s i o r i A - ci - u nn t- ir e tIc i t i - ns Di- i ar e n u - l 1e-ts -cm l r- - -— - - - s-ar t - m o t S ‘ - x ’ u r cl - a I r ma - af e - x p n i - m ’ -cn — umi p oh O x f o rd ft - at - i time L oi- u - r amto r v ml riu ui -- u ’ msu l t lnq ANL Th u - j r r o l e of i s tao necessary sof t’srn - for l’s i- i— u t o i roi’ec-d wit Ii t h e iay I il c - tr s e t i- i k i n g the mma - r i -ca l i i1i’5 -3 r e s o ar c i t o estab lish an d s o f tw a r e N at i e m a t i c s c o r I - t t - r best L a h o ro t o r y C o m p u ta t t o n a l - l i -s i n n - t a l U s a m imi i - m u s t i s - r -aialera of ron - ct iso of the n e t - - -rk so far imai s teen l i b r a r ’ ’ while they assess Lr€- - — - tim - an -I -t ort ’ t h - t- -f m ”’ fl t of tIm ar - u h i t i oim i l n e - n - k mn-i in th e -h-a -’c-I uu i -ms i with -- Oxf ord ap it -- - — roz - - it -- r I 97f- I n i t h i - i s Cr At y itn — n e t - -i- -r a u any real J i r a - nu- an i n C it ’ ’ ‘ - -i - - n - tire -n Ni i ’ t n i t - u l i— kiio- -i lt - I -i c-- _— - -- --- - — - -x’ s-afta- mre -r Lm ’ cr tu r -s J c -t A lso at h am s i-i-cnn - ava l a - l c - o s L -r ANL o n l y r e c e n tl y ‘-l am — ‘ I nc College C- L i r p u t rn -aue -irch a r a aa I i r a ’ at i ii en -a L l e low—cost i nf arn at i -mm The most hum - n t nt g e n e r a l i s e c 1 c s u n - - u n i c a t ion research r e-Lot on l i n e t e x t -love ‘h o of ‘-1 ti he -n im - i ’ ’inu their n-of ’ ii f r om t h e ‘ ‘ i - i tIm ‘ if K on t h e - UCL ‘r ip - ts e f e c t i ’ ol t cc h an - t i - t lie in nm n ted It - C ir e- ‘I’his i -a ’ b- nm or in re-ru imill t i Ic- s on - i r U i r Hi ’ I I i u e - c u n - -m u- i f bai t - i l l ’ - ” i it e d h o st th t- y I intj te -l two—w ai v ir e t o - n - a s— ab c t r e n t - - of syst labor i t ion u_’u u l cm u n ab l e by t h e u s e of in th u - ti n y TI b u f f e r h is — tn - I ‘ou i- ’ e r i t i o n t h e U i i n- on-b a cC of - ‘ L e a ’ i v o n -man - i io n p r i -j e - u’ t n - - is p r o b l e m a’i I ii bc Chapter and tHe is c - ’ n u ’e’ proc essine- sv st i- -- s a- als arc the uievc’ lo t e i r n - in - u m - u ’h m u se T h ey and sy tt- ¾- - t - -- - n-r i is t rii-ut AU PAPi T en - al - ic t h e n - t- - - --1 t - i r ir if ‘r - i t i- mi Un i v - l i lab i is t - u - ir p e i l C- -u 1 at - t’ n - c- unde r sysL i - n - i n r e s po n s i v e i n t - -r mcti’ o t en - s - ‘-Nc ’ sot twi r a- at d c - v e l o p lnq - a im e d - in -i ied c-n-i - Queen of fl S by n m t receive files t - - son- i ft les h ack n- i- ire It is h o sed t h e PPS S d e v e l o p m e n t - - -- - - -- -- _ - - _ ---- - - - - - - —- -- - - - - WT This year - 5 1C h ive sun ul ieci n -cr tware aa am nu r r r of U S u s e r s Th s has been facil atated I-v - P P A N E T a i r Il Ot’ pr- v s h1 J min or updates through norma l terminal tin -c RCA Royal College of Art During the last year the p r irmary use of ARPANET by the Departliest of D e s i g n R e s e a r c h at RCA ha been for research and running are in architecture m a i n t ai n i n a con t a c t with US progr ams of Th e a r e a s of i n t e r e s t n t c r c n t and desig n research interactive ciraphics an d a r t i f i c i a l i n t e l l i g e n c e The host m a c hi n e s used w e r e C MU — I OA Sun-en-- Aim Mit—Multics HARV—1O and Off ice—i The heaviest use was at CM U— lOA involving the BDS system under development by Eastman It has not been p o s s i b l e to im p l e m e n t nia S system in the UK so this use has been a valuable experience Thame s Polytechnic Crowe and Av ison of the Systems Analysis Division Thames Polytechn ic are researching into the implementation of relational data- They have obtained an account -an the- Unix system at RAND-ISP bases and are using the INGRES relational database there ta’ s t u d y the imp le m e n t- t i o n of gain 0 logical data s t r u c t u r e s insight into the problems of of relational theory This will enable thor’s a a p r a c t i c a l i mp l ementation They also have t i e use of a guest account arm the Datacomputer on which t h e y are doing similar work for the pun-- pose of c o m p a r i s o n with INGRES University College London Commun ications Studies Group CSG “ Technology asses smen t of the i nt e r act io n b et w een trav e l a nd telecomm un i cat ions ” is the collaborative project ineinas carried by t h e CSG t h e S t a n f o r d R e s e a r c h I n s t i t u t e and B e l l C a n a d a out All have access to the Office-i machi ne on which a common bib liograp h y Pye of the CSG uses the network for developmas n is being developed and accessing bibliographi es on communication and tel ecomn r amni cat ics _ _ _ _ _ _ _ _ _ - - - - -- -- — ‘ — — -i C n ‘l u s i- m is F L- re -k h-IPDNLT s i n be no t tnt- t ’ c e x i s t e n c e of that is n—a w a na’ m 1- amahu t o b een nsa - Ic clear i n e l v e d I- H t mmot on - i -ur L i n - - i r 3 in Sect m o m m s 9 2 i m p - n t an t of r smnv - mi sers l in k ‘ h i s - - Some of the projec ts t hi t t h e LCD - 9 — 138 concerned but cu p s rm 1 so n- the worl d of research onera1 1—- T h is is not mit u a’onn-tater ’ient ex mrmp l t ti me - u x f or d 1 or v ir d C h m i c a q o 11 1 110 i s i ici —m tc’ nq’’ coll ab iu r— it ion is can -ve rn e - I ture of i -m itt o r between 1 it m A l th o u g h -oide o f n-i i-ur htlv th e mmmd mr s m - m j - d a a t e t h e n e t w o r k t h e - - erk - ‘m s t lie A gai n Nl i c k - e st w e ek s cit d i e m - -x cim l -c in qro-a s turn Ion i - i ’ I L i S 1C hav es taken C d i st t n i - in c e in 1ii 75 1ctivel’ ’ am5 1 n’ -sore N- f i l e - t n - i n n - I A - n f a c il i t y without a- hicn i-s m a ’ TIme t r a n s f e r f a c i l i ty has OS sj t s 1 -- m n as - ia l a b - i i - use of our I- i - - Pen-n Office of ava I l a I - i l i t ’ s of to aml m ’ m-a st r e a l U- s i - n i li -an t er - cr -_ - — - - mis c c work in a n m - -Ce m-d ere the’’ The es s - n i t L i l St r u c - b a s i s of seismic l - r a-vl u ’i siv i-u y seismn - ti 1 -e m n o d o n - i was nor-a’ c o n s t ru c t in -c 1977 w i It era ‘lo n - am -u - - in - er- h av e nm -ide - wlii u ’i - I c t r - u t i - mm o f a - -a r e xc 1 a a n e f i l e s ‘-‘i i t h a’ ’ u 1 d m -n lt t im e Uf users - - u l d not h ive 1 u r a ’ - ’ t- - 1 EPSS i n IUO n- i rchm w i t h the w o n 1 u I — ” — - - -- ill the t iuc - f i l e - i icat inn -s e n- S — i3 ’ — CHAPTER lO r 10 1 FA C Si°IL-E ‘ CjiVt’ ii’S Introduction The present message and t e x t m a n i p u l a t i o n s e r v i c e s o f f e r e d over ARPANET are extremely important for processing textual data input direc t ly into one of the ho s ts on the net w or k Ho wev e r we have long fel t that these services have one element missing tn -cr c are no facilities to deal with material wh ich does not originate in machine readable form and the f a c i l i t i e s imm aterial are poor for r roce ssin non— textual We are investigating the input of documents their transmission ‘- he ir storage an d subsequen t re tr iev a l their manipulation in con junction w it h other files a n d th e i r f i n a l ou t p u t on dev ices d ifferent from those on which they were input This work is n ot a purel y theoretical exercise Kirstein 1976A during Stage 1 were undertaken involving 1975—1 97 6 As reported in specific ISI and ourselves only experiments These experiments were mainly concerned wi t h t he dat a inp ut transformation of facsi nm it c f i les in to a c a n o n i c a l form and their output on different devices The work done dur ing th i s pe ri od re su lted i n a poss ib le but impractica l system However it has provided us with the necessary in f o r m a t ion as to how such syste ms should be designed to meet the requirements Amo ngst others this system had the following shortcomings 1 It was heavily dependent on the UCL PDP-9 and used too much processo r t ime 2 It required considerable manned intervention and was awkward to use 3 Storage and Retrieval aspects were explored only superficiall y Hence the Stage— 2 work was planned to overcome these problems t h e p r e v i o u s work Stage 2 demanded a Systems Model Unlike Our conceptual thinking and how we have attemp ted to realise it are discussed in -fr - — -ffLj - 5 j r Iu t ‘ - - l m n - t o I mLL - m a - ’ u-a -c ’ass oi 15 Dad e is c nma’ - ’o t IR c o m p u tor 1 nmm ni l ti t erms u s e of th in- 0 C m u m ou r u s e Qf itt -ci I u f o r m i t i on a-c- ire iti ra n of Th m- ri - ’i -n - -c - i f - ic ° flO irm t ime A ta l c - n u n o n - - s s s - n cut ta-a- diu t m c ul t is c- u n - t m oi and - i - - l ’ ro n o - ss or st s - 5 _ i tic i t iO i or i n -20 C m i u n- un tb ui i n -- n t i t - r F - ‘ - 1’ ru ’ - time I tie m -mclu - i n ’ - L r -m f f i c i i n- I mum ’’ m c i i a ’ i u t c i ’ n - y n - f r - mm t tariff u i-I f low - -— of m c c n-OS in nm t In to -r ’r c on e - - f- ’ nr sn a i l N- ’d is iimLe ii ‘ am -- -n sin m e i a-un I - ‘ ‘ m u s e a’ a - h i c i n - i d a- ’ it 1— - ten -ac- -n t - e m ’ -Str s a ’t t ’ ’-’ mme s- ’ncm ’ - r i— a t iO u of o n -cal ’ C L I an is t u r u t n - ’ at and f mmd a b - c -- l’ ii ii 5 - c t ijs m i m m t c i lmu- a - i l l ile sc niur - -— t 1 ii to m’ mu la - n r L- it u -b cnn b c-u i m iii in F - ia - - C u t-re wi wi nk mm- - - _ _ ___ i I i c i A ’ sy s I’ ’ StI n - - n i o - r n -i u ’kn l Ti -i t-n - u -I imj 1 - n - j ‘i-c - ui ’ m il Ia ma ’ s b - - i - r i us am s o l u t i o n i at - r dr- i a r us tub u 1r mm- t m ’ -I ’c-ci n m n r muim k C in a n u n u mir l - b u a-i n- of the - - s e m m t y - i l i n - - -d i u- I a l l - -‘ a c n s l n - L r a a 1 r s ‘i h u d-2 Sot k - t - n - en - i n s - k I H i m - l A Y t e r n - m u U ’ - ’ — a a 0L tear n-a Is u t s-u k ‘ - i -be er ’ i c e - maui _ u i 1 the - m c i j - c t u or i t - i ’ - r-f 0 - p u n 1i- -in C - ri room -ri an n u a l -ici e r mu i r it r- - ui-JO’ t ab l e— - i n i ’n Lie q r i ’os i l and ire r i r n - mui ny thin - - - in - n- r of It is a a - ones Ii u ’ mt L Iu - 1 - P et r L ’ J e- f u r’h ’n’ e - t- -n i- ’n me 7 Ii- nit i t i m e r s - - s’ ’ n - t - mrs -- i l l Jn - ’il a-i m ’ h S n’ I r n - m c o n m i u e r n a i k - t e r n - i t s tim - a i other mc I 01 SAY I ten - nu mi I- ’C- t bme - t y -e- s o f lift-rent It of Iu 4 n - a ’ I b m - l b- - re- mim k e n t the PilL a i m - - u - a - ’r ci n-c -n a - re air a cr ibed in t e r a c t i v e l y i n DL i “ h e - t - ’ r - A-cm m-sS Thi c tmi m- ” - - am r r mpic - s e n teul a - -cuj - ic- in - C of aiim - p i - m e rSi- t o r i ’ - - ml tern -i t i ‘cn- m - re 3 li m O 1 t h a f t c -r o n ’ u i u c n ’mac - in - cn n Lv Iqi t r u- - r b n-y e -i - rt -m u-tr’ St orm -H u ’t 1 - m l 1 a - lurc h act iv ity all dr ive n -n u -H it t ir e an-i11 5 i n n - u i en e m a i t i C mm IL t i O l on -i - 10 3 in S e c ti o n of than r i r s h i m i - ’ mm out li r ra ’il nm m H st brmnir - liii — ire- cIis-u sed f-a r th’ s t e r m i n a l 3u - --si mu it Lm a ‘‘ - c -- m-c -c- - mn -l ashL-ca io am u ta’tum f - a ’ s t 1e - h - v i n C mn - C t h e st i t I m r -I i m1 mr e -n - u - i Both our bm - irt tn-rn nec rer n- a n d our long t- r n - i rn - S f - mc in urie - n - e m - -- z S - - ‘im 1 n m m n e Sn - - m t un-S - S c c r m - br -I iii - ii f - i ’’ n o n - t m - -n - r - ’ m isc m I _ _ _ __ s ___ _ _ _ _ I u 1l u -at - in __ _ _ t tori thi s mi n m i i i ’ ’ I n -tm - S c ’j - a ’r s - t i C r u L-ms s c- r a p i o u in still c ’o mm mn n l c- Lii- - ‘ u n m - O - r ‘as tried la i”s it -- tn-i r ii 10 2 switch ed an mn - Cl I rn p I a - i 1 ‘ 1 c-nm __ iim-iu’ Of t mic - e 1’3’a’ e- rly -- u IL Lai m f f ’ COn - ’ - ith n-f s t - ut- - — —--- — — -——- — --— ‘ - - - -—- - - 141 - — - ii- -a”rr it has clearl y shown the importance i t fam c siur ile naxu ely ‘tree factors for 1 Bulk Rate Discounts 2 Off-Peak Rates 3 The importance of integrating the Information Storace and Retrieval Node IR into the network The first two of the above are self evident h-sen with a reasonable compression am full A4 page of text will typica l ly require 2 5 bytes of data Even with optimum pac icet fill this imp l ies a - i mI t u transaction cost of 12c page on TELENET and 17c - ’page on EPSS at ç eak tariffs In both cases call duration oort access and fixed charges are ignored in the above For significant traffic it clearly is in m n c ’r t a n t w h e t h e r the I is integral to the network -i-r considered outside it odse there will be two impositions the transaction charar e lR and one on retrieval in the other case only one In one - one to the Clearly the irn-uportance of this factor depends on the an ti -mi nt -i - f traffic t h e len -el of c h a r g e s f o r I R the ratio of transaction cost to aml duration of fixed costs the cost of mess a-u- onocessino etc The - analysis of the impact of tariffs on this app li- mt ion are so complex that we have started develop ing effects i- computer program to study the This program accepts as input Datrix of Transaction Tariffs Ma t r ix of T ra f f ic Vo l u m e s Vectors of Fixed and Time Charges Terminal IR Message Processing and communication costs all will 10 2 be considered ‘ill This work is still at sin embryonic stage but be discussed more fully in the n ext report The System Overview In principle the system we are striving to devc l op i s shown in F ig 10 The UCL facsimile terminal FAX 1 i_ s connected v i a t h e T I P and a charac ter term i na l i nterface to ARPANE T’ are an information storage and retrieval system Also attached IR a message n-a ’ ml - - a ’ • u __ i s _ • I - us ’ ma ‘ - - - i I - - - u ’ l - - - - ‘ - ‘ ml — i - I ’ I’ H ‘1 I- - I - sb in uc or ‘ - I ‘i- iii i ‘i mm ’ - -mu ’I Si u ’ t i- - ‘ I I ii -i _ — I _ ‘ _ u n - n - - i - ii 1 a- - - t - mm u u i ’ m m I t u i L I tib i l -u- I u t i t t i- n - i - n t or ‘ I -‘ i ‘ 1 0 ii m ui-b t — ‘- l is t no t c -’h — ‘ tie _ mmmcm ’t “ - - - - - n - u - l ‘ — r-h m m u - In - mn - n _ ic - t ot— n-i s- I - i 1- n- eu mm t - n ’ ‘ - I lii - 1 - - - _ - - u a t - ui ’ u t i _ i hO I l oin i - as it - u -‘ - - th e I - - l s i m ’ -ui- ’ m m - ‘ c’ - a’s- r n r i c r em s mOm - n -s ‘ f l i t - ’ r u - I - m m - - ‘ s - s - s m - r n- mm ho i n - i m ’s 1 c I i u n -i - 1- u — tHu u ’I - 5r - n u i 1 0 - - il u n m — I - u ’t s n - - I ‘ N m mm ‘ - - -r u u ‘r u e n a r ’- i ’ nm In 5 - m bL m c- ut u_ _ ‘ - - mr In - I i - 0 - i m m Iu- m - — - i - i n n - b 1 ny - m e n -n - ‘I’ I’ - -v i mm St - -i’ m - 12 - i f lt lc - ‘ n b u-m ’ n - i I I o u -I u - ii I ‘Oi l - -l - _ I- 1 - H mu r t ‘ut iu ‘m r - I - - u 1 i a - - - ui - r n I I u-i - rm u i _ - ‘fl - - w - - -mu ‘ - - - - - - lie - I mmii r - -I- v a-c - - n- m ‘ - -i - S - - C e -a I ‘ c a - mm - i n - u ’ - Cu V i s ’ 5 smii - muIu _- ’ m mi mu is ml n c’ rmo m e - - ‘ - -- “ ic u- - - Li in - - - - u - I - ’ - - ‘-1 -- - - u _ r - mu u - - ‘_ — - am u 0 It ’ ’ i-mm - -c - t i m- s - - m- - no _ - _ “ i m t o r m i t - ’ r 1 1 - S - -‘ • - - ‘ - ‘1’ ‘ u - i i ’ I - - I - ‘‘ u - - u - • - IS - - f t n- -r - I ‘ -a r m ii - P t - - -r us - In - at _ a - L u - - - ’ - r a - ‘ n m is ’ 1 S i - m u 1mn - m I m _ — I -ma a- u r - ln - - _ 1 - n rli m - i tI - m - lit I Ii- ’ u m H ’ o __ t lm u u n _ em- m a - s’ ‘I e - - n - ‘ Ui - m u _ - t O - L m - i s - m t - ‘ u n- Li S i n 1 n - - -C nm ’ ’ a a - -tm - it-m us a S - - - I iC - i u- I i - - - - a - i La i t n - t ’ 1 —‘ ” l uma I ci LOS en m- ’ ’ -u “ Ii’ i mm Oil f ml- O - - - u’ - - t r I s - ’a r mm ‘ u — t -_i_ t - - ‘ b h’ - - r -- r rc -mm u m ui- mn ’ ni ‘c o n Sn - - - t i u i - in-- - - 1 1 - I - - n-n- I 5 0’ i u t l u t - - I u on- H - - -r Ii mc’ rfu - - 11- at I i n u - - mm i c i-c i - -n - tcim imi - - m tn -u ’ F - - I -e - _ I - - 1 L - m I u_ n - n-I - - - mm - mm u - - m - n mm - C s - - - i t -- a - - - i k- f - - ii - SU F-a - - - - - - - _ - - - 1 -t i c —- ‘ tuu - — - -- ‘ -_ ——- -- — cc ’- - ’L Lola 2 u’ a ir i n -u I — _ I _ - -- - -‘mm -- - __r - - - I - - I 5_5_ - ‘ ‘ - - -_ — i ‘ _t - _ -- - “5 _ _ 5 Nu ‘I - I m m _- - 0 - a -5- - I- -- I ‘ - - - - s __ _ j _ - ‘_ -- 1” — I - j 1 I _ _ _ _ —- _ - _ - _ - -- I £ ic - u - m r - r u - mI L u c ” ’ - ‘ r im u n e ’n ’ i u s i v u - l ’ s- cm ’1 e mr mu - n t m n mbe- a-cams ft to scm ’mmc’e i rn -- n- n - u n - l u - i t l m au1-J in n - - n _ mm m-Vcfl ‘ t - 1 ca - en -ta lf v N et -n - - uh ‘ Se- n-- mmn - c m ‘Ffl I I u u 3 I it aaa t - t- ’ - - I I c- u ‘ i n u in - c - - c - n c ’ I s - ’- t I mu I I I a ‘ -u 1’ ‘ - Hu - n_usc I I t a - ru m - the tu t — O r - m n I- - This tim O DC Ohs NI’ cl _ m t - u by t o j iP 1mamths a - --n thu ‘ n-an- n-s n -ut s o f n _ nc- - l a ’ t m m u ’ - ‘ ‘ i l 1 i-fe- lat I i n s- - u cim ’ i t m a - iL ‘l 0 1 - n - t n - u— -r q clu se t b ’ t H-tI a- ’ u c m I_ i e n_ s m PDPu ma ti loo n hi i n - u - r I S n - me t r-e s i n t mm e ha- ‘ s-t e ’nl Iu ’iu- -r m - i u t 1 n u I th u i f r ’ -m I — s -i uui oi l ar m - i cmi ’ - t rim in - n - in r i s i ii h’ i ‘n-cr jbe a iii ‘n - mm ’ uu ’ -u b Iy a CiL S Is n-v ion lot U s ’ i an a l l n -v S i- c-I fliO n i-v u in 1y s ’u i 1 1 m - ’ mu - n t - o m -u n h - the s y s - t s -r l iii’ t o m m l i n -n - 1 m m mm U t e- -- siuuni do cmmr - -iut i- m r m I I - u i ’s- m I lie m i u s r o l u r o c i - 1n - n - u - l- R -irn m s’ru t - it u ser DIalogue puurt t ’i ’r I - ui i m a Ls of - i ’ - - ii ’ s c r i b m s ’ u l u ’ - I n i -I -i i ’ lii - su n ‘5 1 13 - an -i 0 3 2 5 - I - t i - - mu u ul u - in - 10 3 3 bu ‘ m m - ’ - i t -m - t i lu u iii I F mm - m u f l u i c - e a - il u ’ - n - ‘i n - u - - - iku- ‘ m - ’ m u - ’ uc am n ’ i i i - 1 I u’ L -• t ion a I a 53s i i I n-- ri I n’ l id 1 - mn - su m - I t u ’r V i - i’ m ’ ir’ m n u l o r - n r t e - -m i - -ut n - bu ic -li and h i th the - - tn-N t i - Fam n ct iu u i a s 3 ‘ ’ ’ eS Lt i c - ni - I n - c n m n - u ’ o’na- ’ e’-r n - m-I L u c s r r - I e tin -n -n the- cc-n -ne -ct t O f i S r i t h - - n t n - u - ave at -_ s -u t iu uil’ ly open - ‘ -- n - m a - n - t m - e m s to t ie - “ X GP t lnu _ ’ Nm ’ N of 11 7 n- mu n ie - t c n n a i c a 1 I- ml u ’u L’ n-’ n - 5 - m sh i c n -i’ j 1 u m im m i U e - b e -n - an - ni u n - s r sn-- n L a n - i an the XGP in-i s mu - i - i l i c- u t i o n i’I i P ’ and USe j r - u n - been uneu -r - m u J e n -’t s u mj iea La o -1 h t s n - I f a i e ’S a n o h a s hoe-n ns u uI L i i i ’ L cm o q - m Cs’- Som i I t i - m i t u t j s j m n - i r u aunt hi Cu t e hi ’ 1 ci i c r IR i n - - em N ’C the- f m ’ u n g er m miflaectl a r - l m ’ i t c to u I r i v c o ’ - ’u ml I u i t iU 2 5 uiiflu 11 5 1 - - n t m u l l Lri Lii-- C e - r i m — - S i n an - a - t I me ’ ‘l’he a ’ ster n_ s n- -si i- s u ’ a- t i n - - 11esa a - n - u i c - ‘ 1’ ice s- s o n m t lmn - m I - ‘ u s - ruin u ’ s t - u F u i i s -h -m un - -i-a u ’rmi l ie a m a in-n- ’ i t ry I c u mu -n- u S m e - n - S ‘ - h ’i - m i m j La U ’ I - si l l s- f a r m c t ion s tn -si L u i - 5 0 5 sire f- u In - - i ii ‘os51i lC in - It w i l l prn- -batH- Lu n - m i m a t e -un 1’ 7 A m I’b i’ ic - ma - — l ’ I C ’ l I co imf iq ur - it con bu -ca use by - i a ’ Liii n - s u n - i c of n- mt “Ih- 2 u m iiu tn -n - a 1 t l - auto-n site in - ma r ’ c -C I r e -u i e ’ah ’ u t a n - t r o T le d Co I ’u - - iu Iv tn- u mu ’ I - u i i j -n ’ -v i ou n l 1 ’ s iC ‘- ia-i ’- U s - f- u r n - i t iii s n_ - s t e - n i -i t O u JC - ‘- ‘me S_ - ‘ m n n _- - - _ j n - ’ s t u ’ n- t t - a n s nr i s m i on hi ’ u ’t i -r i mu cerime d 1 0 3 - I w i t Ii t mu ’ _ - _ - — J45 — r—- - - n-n I S ‘ ‘mm n_mm -a n__i -‘ uI- i - H ut n‘ u ‘i-I - _ -—S i -m 4 - i_ _i I- ’ ‘- S - _ i_ a - - _ is I 1 -1 r - ‘1 - ‘ - - S 1 I - —- i H I - I • j - I - - - - Si __ I - -c -n-nmpes it a -ma ‘ c-nt ‘re-ss i- e - to the MSG mess- uq e system tIme of u b nu 1 j ru m o f t i m e - _ ‘ I tn -k s es t- t h e D at a c om p u t e r and t h e supe r v i s i o n to both is almost transparent to the n - I m u s s - ay e t r a n s m i s s io n mn -er T i- retrieval rn-m al e can have un -ne or two stages Ime ’t i t l en-I t ion conies vi u mm u t i f i c a t ren-’ u i v m Curing stn uiutht ‘ l Ime - t c-Lmlilf la l Su n- 1 O s n- cmaa - - s b m a t this ‘mm — f u c s i i n i n _ l e d o cm mur -n - nt n - -I We ruin l ‘is a I - r n - i a - C text retrieval s i mp l e r e i i i t me m e s sa ge thin- u s e r may from a conventional t h e n be a r et r i e v a l m u s t n_’ t 1tc rru ative 1 the user ma n- access v i a a l u i u U e tlm G vi a t ime facsimile terminal In I f i c a t ion and r e tr i e v a l w i l l e l-m u - ’t n-vm ce to n-- u nu ’ r i t e h ive ’ t hey am — ’ a ll to ‘n-hi- -Ic lists of s y s t e m MSG mn - of the availability of facsimile files exe r- ’m se c isc- n-m oe ms mt i text t ue Because the be an inte t rated al l the facilities aV iIab c ho ’ s ’ u - a’c- r to fomma - ard rio ti t ac ationS u ’ iul m e ’ssn--s answe r messaqes and simi b r fu n ct i o n s The Fac s i m i l e T e n nj n ul 10 3 2 Prior Lu a dcn -crip t i n - u -t i m p l e me n t a t i o n f a typical users dialou iu a’ in our an ove rview will be given mod el The facsimile termin al t h e d e t a i l s n -if o u r u F A X I is sketched in F’ iq 10 3 Physically it consists of a Plesscy 4 6 m i n u t e - analogue facsir ile FAC ui via a simple A D 2 level cm n - im v e r te r to a 2 4 K 8 b i t s or c i IN T EL5 - 0 80 rnl -m - u ip r o c c s s or Also attached at p r o s e- n t I S a f l o p p y d i n - k a keyboard t e r m i n a l a n d t n - c asynchronous cnsmnnels - ire a t t n -n - lied to the T I P and one to t h e de v i c e at t a c u i ’ - pP bec use we ii uv u ’ hi -ad some temporary software problems in mu lti p lexinu t n - c streams ove r the PDP9 iuP interface but this PPDP 9 — - will be re-me -died early in 1 1 7 7 were used because we’ p I inn emJ Two asynch ronous fo rmat channels to enter via the TIP terminal V rts t h e s e w o u l d c n i y s n _ m fl pc u rt t h a t f o r m a t a n d w o u l d o n l y h a v e one v i r t e n l u ” u n - n c i tj o n uv e r each p o r t W i t our abandonment - - n - o u l d now i u and n n - I s o t - m T h i s we w i l l shows in - it n t ’n _ u r a l t o m u l t ip l o x I P now c o n n e c t s s t r a i g h t t o t l c PDP 9 p n ’- s e n L it ise n_i s i n q i c s y n c h r o n o u s i n t e r f a c e - -s c on is i-vu ’n t u a l l ’ 1’ t h e • imi- i La i’ iml t e x t used and f a c s i m i l e d a t a C o n c e p t u a l l y t h e r e f o r e Fi g 10 3 p I •mnne d con ‘ i qu r s i _ - t au-mm --- 5 — -- -- - ‘ t h i s a p p ro ac h t h e - - of - - - -- -- ‘5 - — ‘— ‘ -5- ‘ 5 5 — n 5 - -‘ - ’ u 5 ’- I I -I -mm C L - - - _5 - - --- - - - ‘ ‘ -- - _ - _ - - W I - 14 8 - The standard 4 b minute analogue facsimile devices must first synchronisc wit - 1 each other wh ich ta k e s up to 1 5 seco nd s and then read a full page at a time without interruption It is impossible to guarantee any specific data throughput via a packet switched network like ARPANET th e r e f o r e d a t a is s t a g ed always via a floppy disk At a later stage we can plan to replace the analogue facsimile device by a digital one in which case the data flow can be interrupted after each scan line In this case mu floppy disk should no longer be n e c e s s a r y 1-lowever a backing store m ay still be desirable for other reasons see Section 10 5 2 As described Kirstein 1976A the analogue data is passed through a simple threshold detector sampled at about 2 4K bps and the data output -is a synchronous series of eight bit bytes though with asynchrun un-as format start bit eight data bits stop b i t T h i s a l l o w s s t a n d a r d c o m m u n i c a t i o n s i n t e r f a c e s to be used between the FAC and the j1P A synchronous data adaptor w i l l be used eventually instead of the asynchronous one This is more appropriate for the usual situation where the terminal is remote from the network and must use medium speed data trans- mission facilities The keybo ard is needed to control the system and add a d d r e s s i n g and document n a m i n g f a c i l i t i e s f o r document t r a n s m i s s i o n and retrieval In our case a standard teleprinter is used operat ional systerm n_i simple keyboard and cheap 20 character wide printer would be quite 10 3 3 In an ide - lucite The User Dialogue f o r Document Transmission A typ ical user d i a u u u 4 u c shown in Fig 10 4 bel ow 1 for transmitting a 5 pag e doc ument is A commentary on this dialogue is given Start-up The facsimile system resides on a floppy-disk and it is load ed i nto t he memory b y typ ing an ‘L’ on the user console It then initi ali ses itself and prints out the title and L “S S’ 5_ _ _ _ — - _ - - _ D iALO GUE $ COMNENTS L CR UCL FAX-MESSAGE LSTEM VERS ION X 04 YOUR NAME x 1i S’u I ç R 1 Log --n XYZ -CR PASSWORD LOGIN 0 1 TYPE CR FOR HELP 2 d ional User 3 Addressing and naming the document 4 Document Fe- d i ng 5 This message is composed 6 Sending the message 8 D e l i v e r y confirm ation of facsimile da t a $C CRY COMPOSE A FAX Mk SSAGE FILE TO CC NAME KIRSTEIN at ISIA CR YIL M A Z a t FAX DOC l B CR -CCR 5 CR INSERT TIl E FAX DOCUMENTS AND TYPE YOUR NUMBER OF PAGES MESSAGE THIS IS A TEST MESSAGE SENT BY FACSIMILE S Y S TF t - AT tJc L CR 1EXT IS O K KIRSTEIN AT ISIA Y1LMA Z AT BBNE FAX IS O K 7 Deliver ing confirmat ion of t e x t Fig 10 4 Dialogue for sending a document Characters i nput by user are u n ie r limn - -1 Carriage Returns indicated t ’ - CR - - - - _ _ _ _ _ _ _ _ P - ’ --5- — 150 — the cu r rent vers ion number and a s k s th e u s e r h i s name and pa sswo rd to p erform the log in procedure I f the log in is succes sf u l it pr in ts out t h e FA XSY S prom pt char ac ters and waits for a command input as shown below For subs equen t fac simi le op era ti ons ‘ — ‘ this portion of the dialogue will not be repeated 2 If Hel p a C R is typed a H e l p f i l e is produced on t h e con s o l e giving information on the use of FAXSYS commands 3 Addressing T h i s i n d i c a t e s t h e a d d r e s s e s of no ti f ica t ion abou t t h e availability of the document and the short text message The Document ’ s na m e is c h o s e n to be in f o r m a t ive it will be printed out later for the recip ient but i s not used to identif y the document see Section 10 3 5 The iden ti f icat ion number is composed at thi s ti me so t h a t the comp lete sub jec t f i eld for MS G i s forme d 4 Document Feeding At t h i s p o i n t t h e f a c s im ile document is in ser ted in to the machine for scanning W h i l e t h e t e x t m e s s a g e is being typed in the scanned data is processed and stored on the floppy-disk 5 Me ssage Composition The FAXSYS provides three editing characters which may be used to correct errors during this period tA to d e l e t e the l a s t character 1 W 1 N to delete the current line expunge message -— -- -- - -- -- ‘S These are 1 t- - a rid dcl - u f fa rn- amt u b e-’ a ce i- in t I r -Acro ’ S -ssn --r mt S i - i- i ’ i -rn - - “ m e - ‘I - c ’ n - 1 t i charava or ce -mm -i i n - n - n - 1 - c -a rm r - m -’mm c to d ’n- - m - suit - thcuq i t n -t C a ‘i f - m’h i i - a mm_ i n -n - c t m-m ’ - - m aa mutt m A - a _i_i o l k u ’ ei ‘ un - be mu ed rim - - -at c iv n-d the ’ _ i t i r m mes s moe m r - ì Ii esia s a n - i- ’ n -i_ i n- ‘ m n -C coimsia ’ - m-n-i t im e - mm 1 - i mb - lu t% ’ — hu e ‘ u S su r n - S ’ J st c - ’ mbC u t_i 1 - a n- - c-mm c-i - n-nc are mmcc C I u- La m i ‘ibm - -n - - t n ii 0 r n-- - m n t u - mm t i i i iimi- ’oin -t ’ - 7 Ye n- I s n - ra- the ir a c -tm ’ 8 u mmmii m cmi 7or -a r t ont Va - t in - -n - ’ ‘-mic a n d t l a _io cii as non— C d n v-cry ‘ ueucmei dc-livery l xt n - mi C ’ I- - i c ‘u ic- from - n-sc- i’cm- n - the UP L ime riil n-m t u ’ t r ae - v o l uses t I m e - -mst n - ilt- t a c s r m m is - m t Lmm _ Thus m ’etr ie-ve d I -ia any al p h ir mim ’m e - r m c to rn ’ unal p o r n - i n-rn in - tme - - t -r u€’ i O ’ rn - -- mm _ i sun--h - tr -m rsm n tt -mn - i r Vn - 5 a 1 t 1 C ’ S On n- 11 c e r n - ci mcn - ’ t i n - fr’mr m n-c- fi c- d’--1 e r - sc- _m m c t h is i ti c- s mc ’n- - sful s m - _ i Lc-r as in tem ’ -n - c-Len -I Ui mu C h t r - m f l m t ’ m S m - o - m i e ’ J e- n- 0 n-c m0 -od m i t r e - ’ i n - nm mm iv - - a - ri n --n -I n-lit ’ system it m e um C m-i t The Dia u - - u b in ce t h e n- c c j - r e s -q - c s n - r - 1 ‘ - - IIILS 1 m m rn- 10 3 4 uye - - m 11 1 t m- ’’- i ’ m-c - n- m’ s tic i- 4 Cm s im a - - - n - mn cart be n-’ fli f i t_ i - ’ i - mm m_ a nmm s-s t o m -1e -sm o e cu r t t n - i ’ u n- ‘ mt ‘ m t nt of fli -55 mut- me s mn -q- s ca Cn -un -n -nc of his work an the am en- rn-il this case’ tams dici Lu gu e n -m ay well Lol LOW I m ‘ ’ 1 m d m n v - -mn ruq r- m mm -5 t ter loyyLny mm tn ‘AXUOC I of ic -ru n - ti C -c a ’ s Is w a i L i n - header has b - ’ - i - vi ne -i f i c -ni Uhe- on time - Tenex whici- access bm - d - ‘ I i mc h-s a s i mm for im-m 1-C C-n -mi N A ’ h I 01 han m ’ n-t mu ’ 1 F q ‘ - SA and In 1 0 5 that This the length input at p a l s - C 3 - i i i - m g 10 4 The n aim m ul - r in m u a r a n t h e r i e s is a Dat - i - - - - in s- re ference t ’ I ic file Us m 1 ’ s t i l l W i S h to 5 - —- — - — I 1 s r — - S ‘ — 1 in- — d n - CR T ry i n g i_ i IS I—TL NIiX LOGI 1 3 4 10 K I R S TE 1N lSl— SYS TIiN —A PA S S W O R D JOB 32 ON TTY 11 PREVIO ’S 1mm — i N U U II AVE -iSe — Ni UC 10 I 4 V cr s a on 2m 4 I c -h JAN 77 I ’ 20 JAN 77 04 31 if n- n- 1’ READ 4 Jon 77 0 4 3mu 2 5 O - m a-n - t—ex om m m in cu mscjs c’ - sts CR j e 2 9 -1 l from iSi u re-va ut hate 10 Jan 1977 u r n - n i Kirst a- in at SR 1—Al Subject I ’m cc Tin- ’ D o c u n u e n t u I R STE I N VILMA S 1 5 4 1 0 ACCOUNT 1 c-ri 1 I ’1 RI U TE I N AT t IA —— EXI -IC — ‘ Ilna - 28 O f t Y m e s y 10—Jan—77 I” S - m uU Docum ent a v a i l ab l e 5 at 1- ’A N D CC 29 ba - meet p u q e - S i-an- 10 CCA 7702100 253 AT in - IC AT I3BbL -CS 77021009255 D i a l o gu e- f o r Cli i racter - is 28 -Cask pages is o v a i l a i le at IT t- Y J’p U OF ENGTII 5 PAGES IS A TEXT m-I LSSA ’Ci-m SENT BY TIlE I-’A C S I MI L E THIS m n - ay e s 0952 THE FACSIMILE DOCUI’ILN’T NAMI-1 IS AVA I LABLE AT C C A hb LU S T L D C CC 7 7 0 2 1 0 0 9 2 5 5 u c ce s sin y input S Y S T E M AT BE UCL the Message Processor b ’ mim s ’i is under 1 i n e - I — — —- _ _ _ _ _ _ _ _ _ _ 5- — see if t an- u 1 - ‘ t i c s te -j - s of t h e ms ’ u’ u t c r s m 1’ 15 5 m i u I 1m cr - oeI _ a ‘ _ -l ‘ -w ‘ n-s a rn -nm cn- 1 1’ ’ J - - ma i d e n - - i e - r s t a m irt n - u l_ S m- - m - — m - - t in - u Ia sin- I- ti -m — - 5 - - -- — - -_ - - --5 ---- _ n-il tim s an-n- ba a ncin a tn - c- loe -n- m - t un-m ’athe iu ci aI n- In- erm - tn -‘r t e - ’ed ama A t thin-- an -a -n- tIm - m a i r c i— cali aI n-mia se 3 a-a 1 - i 10 -i _ c amaf m c i l i t a p rn i ce ss inui mm for n m can al e e a’ i e m -Ji mm Ie -V c- the cxi r e L r _ n- - € t ’rm -ian u l a t i n- c i _ i C - i I m r e ’ t - ’ 1-a’ n - I c-f n-n- ac - iU- - Lha- a a i Ib i s S 2 mm mU- n- Lfl ri i s - m a i ’ s n- s m l ba C-fl 3 - a- ale- ts r m c cISC ‘‘ a m n -ni t F’ can-c-- he 4 i im -‘ C -m Tl i On ‘ C ’ i Ole io ’- c n ot _ - l i a mn -h ut it ma ’ -_ - -- - - - - - ‘ - - t i n - ’ m l i i t ’ ’ cam mm C 3 tn - ia- t’ “a- r’ ’ - L-- I Tin - _ i I C ‘h” an - ” an -u I i’ ’ th rt -ucmrm r ef l i v -mc to tn-c - mu tom m mt ammn - i c - i c - t a ’ s_ mt- ’ n-iel en-I t_ a ‘ — ‘1’ n - t - m mr a Li IS i t ic- I i ‘ h ’m t i e ’ rot --‘ t i’ n - n - - I n - m i — mm IS IA file ’ in t h e tn -s Its ca r’n tn-C -n -f a mla t acum It CCA uS a r ma t a - n mt_ 1 o t rnm c c ii 1I o 5 I mu I S m ’t m a ccci Es - it m- - 1SI “ a ic f i f e 2 i lo be r e - t r a c - v e c docj mu ’n -t ann - ‘‘ s h r n - h n - i s o - - f m t a i m a s t mm - p c _ i n ’ c - c ir u ’- -N - 1 - ’ 77J c-r’ r_ ia i ‘n-fl’ - ’ - In- ‘ ‘ t ’ u ’ m- u a - - 1 “- u i— n - i i i1 at - IO n - Ft in - IA can tO ’ L ’’ tn - Cs -ui I run - 1 n-’’ m m- MI-- I lm ’- th’1 m cc-- I n -’ - m l a c - S I S fac i 3 he _ ‘c “ I- ’’ a1 COflhill Unl c- U _ in- i mn f - n - a n m i e ‘Li’ ac- a n -u ’ ’ ’ ’_ i ’ S- ’mi L _ iui m rm d ma y n- - a - n n ’ - Ou _ ’ ’l l i ’ ” ‘ 1 t A n- 1 - a aol Iem es -- l n - - i Spec m -ems n 1 a’’ o- u ’- a m ’ n -u cs m’miIe cmic on- i a 1 In - n - - e- -n - u t --n - - ‘ic u-rchi’ - Ii - n-n - n - - ha as -‘ - to am - IC -r - m - n- _ i i a c ‘- I-cl e u as 3 I - I L toa ma of -n-n- j 1 0 1 Arm _ - x m - I - ‘u i - i t n o n c-f t I m fe-- m m - mu-h m c - mm msL t m- nd a f - i -r j mil ’ fl_ I- -m i n - c u t -n ils - a 4 i s lin k - u n - ui t ea -intl also tn- sc n-n-rrc--i_ -t If P hmi d i - I _ _ _ -_ _ ‘ $L CR UCL FAX-MI -I SSAGL SYSTEM VERSION YOUR NAME PASSWORD K IR S TE I N XYZ 4 CR C R LOIn-IN O K M CR ‘-In-U L a s t Road V 1-R S IUN 01-’ 10 J an 1977 A 1’RIL 1 197 u 2 old msq 28 d i sk Pan -es $4R 29 CR TRY IN KIRSTEIN FAXDOC 5 77 210092551 CON1 ’IRYm j CR • 11 tX Ml sm Ac I pEC l-m IVED $L C1-t Fig 10 n - Dialogue for Retrieving Facsimile Files from the Datacomputer _ _ _ _ _ _ — -5 _ - ’— ---’ ’ - 5 - ’ - - - _ -- — c- u_ a n - s r m n- 0 Ii — - ’ 1’ St a g e s n --f - by Si - t- n -m -mi mm - “ r - a i ’ it f e 1 1 3 5 Dc-i i c - i am 5 ‘ ‘ lu nfm- mn- s j c- m i-cu is ‘‘ - ’ ‘-c - _ i n -n c- - n -n -t 1 _ 15 -- IIiCuVs ” t t m1 ’ t tm ac - — 1 a Lu n- 0 ’ S -u-mu - h ’ ss m cd I b -n- - - - - — ‘ ‘ - m n - 1 1 0 - - ‘ - ‘rn in n - i m_ bI n - is n e - s n - i - - es _ n- S • n - c- n-h e f - _ a m - i t — cc an - i- I-c 1 0 n - na u n - m a t u 5n - — - — rn - a b c- -_ - ‘mi ‘5 - i m f l - - n - ’ S - ‘ a - ra - - ’ n-a k t 0 - n- m ’ ’ C Cs e g i5- - ”n - -- ’ aI in - -— Cn- - j C’ — t I C ‘- i n-A -- T s’m - cm a - - t m ’- ’ -i a i” - ii-hI i 5- - - ‘ mc m--sm ’ te I_ us ’st’- n i a l el 1 n- n -’ — Fi ‘ m ’ ti I’ — - I u i c- f St n - m i ’ ’ mtmu tn- m n e ‘Tn -i - l- ma - 4’ n - n - u n - I l I U 1 n - t’ f i n t m mn - - ‘-fl _in- ’ - ’i ‘ t_ -n -’-c - a ’m l in k f -n -n-un -n- sin 1 - 1 ’s ’ ‘ ‘ me ’ ic n i n--fl G€’’ ’u ’r- t i ç r T b- I_ l i e - m ad _ I t cia -n - ui - n - be- n In-mJ ac- s Lm a i in - t- m r - t m - _ 1 ‘c- ‘ j -- f r—m n - — - - _ rr- n- - m m a Im - ’ m c - s - a’ 1 5 f l ’ n - - miim ’ mi ‘ a a f - c ’ a s ca t - - ni a Fe-s r t ic - in - C u mai m -m e’ a 77n - ” I al u 12 5 5 1 it is c m-n - is-nt l a l mmmc c mm o l e di n-l u’ At --n- in u n m mu - t m - L at r ic-vs u-_’ -3 - mm-m ’ m d ’ ’ ice t u r n ’ cad m ni - aces v ‘l’h s -n-- J - ‘ ‘ i I L m - u dmI- n-I in - L u th e - ts - ’ut l w - S c a n- c - iaa’ -a 1e-r S rci nm-n 1i si -- ’- ‘l’hfs c’u ’ d c’ t is r- ’ i i ’ -u -c m a’’ - m c—c - i mn -i s _ m a - ‘c sir-’ le 1 i- IST IN t o ‘st -re t’ i corra’s a n - u 1 niie -a t t ime K Irs ’s s_ i d ii -t ru ’-’ ’ t he r n - i c - i n - i on n--h cont ‘A C IbI u i t - ms e eaI _ ni in - m s n-n-a - mn -nt r ’- i 151 fl er r t u rm -5- - -- 1C m 3 t h e as i-n-i - -ma st’ mt - c -i IN 77021 C I ’1 92 551 ‘5 - --- --- n -n t h a t m e u ” m l’ — If site password a t a Set s i t e s d c - s ir e t o be ’ p a r t i c u l a r l y se c ur e ti m e - a ’ I-i n-n-- u s u ’u i l’ n - t h a t p o i n t in t h e p a t h c - n i p time u s e r s w i l l be g iven password protection by a u s e r K I R S T E I N a file FAXDOC Itic -sC wh o s-e MAILBOX I n - n - G I N 1-’A L ’ SI M J L E S 1- eIRE ’I’ KIRS’ i’J IIN If we desire a’xtra c-m e cur i t v so t in -mt the ti me € ‘ n - U i V u i l C f l t con i a nd n-- st ora il y a - i ll w o ul d be 1 uml I N 1 ‘ACJ M I L E aLOG1N to tE a a al alr em- n - c - c - u ci t m-iul rcsse n - - w i l l la t e r to the file via b u s In- 4 S t o r i -j e I m d on c c n t r u i i s od i i i ty for i m r ’ --d l ’ r n - n - - i f ion ‘in- It n -c-pain-i t o - m t f o r ar c h i va l s t o re ’ of ian -- u i— s -1 m - lie- ’ t Si u r i n - I fit i f l a ’X - ‘1’ t s ’ p m ‘ia r u u s e-m n - o r s and u s c - m I n-s a in- t m j i efin - i V Sliec iul am Thus ii c- l a si ’ui a uc - ’ ’ I m - u in Ea l mm n— 1 in-c In-i’ i u t i n - c - el ‘l’hie system is in-n--n- m ‘p I I b j I _ c- mc Wi m i I am L i i i c - - s n t m j n s o f t w a re a r c - n - u n - elici t - il hardware - be v i e w e d c an othu - r t for i inu p l c n m en t e d u ’ n - i m m u u t i f l J dos -nm it is i ma t en d o d -a t _ a l i t _ a’ i’ uu ie st in -tm m an aujernent services - inn - f o r pr o ’ a d ha ’ tEn-- ‘l ’ h u c- to h is fac simil ’ t e n - m i rn-m i l a i g e— s ca l e r i ’ i m v i nq n - a t u Uit - u—l ’ ingm m d t t i c r c- a tC ’aI C O N N I c ’’i’ sy s t em sm t s ’ i i i I i ls ’ ‘ 1 lcd uur eX ‘ ‘ s r ml f or s’ -i m mmni - - m-m m ’sr u p r i V i l en - n - c ‘-- i l l RI - lAD i fi i1 ’ j _im E o n - i o n - i l - ’ S lit - m y’ t s- nm r’ u n - u n - - c - s - n t n- I a i I i m m e i a ma r n 1- n - - l mm CIs m ma n - i n 11 and l r u r a b n - u r c in cxcs-l l e n t i ’ n - x mi mn - n - I lame - a l for our ircccnt S In S c - o n - i o n 10 1 2 U Ti ’’ 5’hie Dn -t -u co imi m ’ u I _ - m-r ’ i n-nd c or n p r i m e s existence of a file l a ’ s c-n t tEn -- Da tn -comput er dai t u mmmd SECRET • CCA is a i m -n - m’ a ’d A l- mI’ANi- ’I’ f- m c n-ui t Archiva l v a r i o u s mi s ’ t 11o rIm h o s t mm storau Je - mUle Us- st - ira n - a - n -atm n- -r - p u t c -r s m I _ lie ’ C h i r i o t o - n - i S t i o n - of L f e r i r g a Iso ‘F A X n - ui- ic - ‘n-nc- Data o r u nmj u ut cr I _ l i t - ba it and SITE lAX SITE KIRS’I’EIII in - IA not ification of t i m ’ urn addressee ’ s m a il la m-n n - c -r u f a a ’ ii i Ic- syst on h a v e s protection pass’ n-ou a lms the site ISIA and I”AXDOC - itself 1 0 ‘1 is a t I S I A an d password is SECRET w o u l d be LOGIN a Thus to access - mm na - c e r t ai n — in t h e - i n - i - m u - mm - m s i -itace nS I ut n -u ar id - ‘ — 5 b r i e fl y LI m - - way --—- 5— - - the d a t i files ‘ — - - cm I - - s t r u - t u r s a s set u i ’u ’e ’ m s c d _ - ‘ - ‘-- ‘ _ — — — - --- _ ‘—_ -_ — In - - -‘ - - — — ri-mn -es c- am S n - e d t to AT’PACII em - - ‘ LC mIN ha- ‘riri t si - i u Uq j m a Cc-- U in-i -_ n-c r - S - i l U i ’ D at a— I i ’s ‘ i i r U n - n - a ’ v t - ‘s - in -‘u mr- n-c- e ut - t_ l t’ it mc n n e - - - n - u n - S u L ’ i e it - ‘ - m a im i m l mt mid ‘ ic-n - W S’ in- iS I f rn-c m s e t m- anu u ‘n--n- as ‘ a a n- sc - ia i’ -m - c - w u u r am pr a n-_ ec ’si- ’ m it - v u u I e b e- pa s n - ib n - e th is n -I i n- I m id r -n - sc s ‘as then ran - i ’ a y in -a ’ mai n - L u S_ he to -jive a m n -un -n - u m r e J t c-r degree- of -n - i it tIm e p e n - v is j u n - m s u se 1- n-CS NILE diuuv c or -’ cml t i e ma - a t a f i c i t i a ’ u m a box S i t e - intc- t ime f o r m e d ‘ ith -n- m n - I to fl - c- m u m i r a direct in- but he cum ul i — C m Un-m n- p - U c i sn - u o r r c m u u m a c j m a - m f l s - m ’ ’ n- n -n c User fin-rn-a The 1’ t_ ’Si Fiia u m i r e - c i m u r ’ ’ on t i m e D a t n - c m n m m - L er i r s a t r e e w i t h ’ f ’A C S l ’nL - i i i ’ a t tim e t i s i t e m a c - a l e s mna i c- r l- ’ iu tn-i N I — i user ’ - mis t i n - - p - u m - w u ’r u S km m ’ ” i -I a - c s - s 5 m i a ’ l c n - _i Wi th - m ama -c m n-s i_c- - f mc L r e ’c n-m liV emam ’s t i e ‘ s-mn ‘-i ’ c-c’ ’S rl R bAI h’i’t_i j -a n -n - - - - i’ n - m e- to I baa n-n i m on- n-c iv i ln-’- c aedresses I ic pi ma’mucm’ -’ a’ -a rm e -1 - c e - n - s - mrD ’ L 1 I c - in t - s t i l l li e ’ tIma - c mr W in - - be n-ole If m r i’’i ic-c Al a m d na -i c maro ’’ s ’e - - - r n -L Sc- a n ’ s u e ‘i ’ i is mm n -’i u ita ’ m - a n ’ ’ - mm m v i - e -’ - p u Ssw O m S Vi I in - a c- -m n - i ml L c-mn -u r_ a s e - m - n -- oma s u e ‘ m s e mu -e -t mn - tm to a mu — Ic - Iii e imc - u h a specific p ut n-n 1 ’ n -n - m is - I t r i - cc mn - n- ’- n - rn - u 1 n -m ’n-’ n s ic- a f n-u mm n - is n a - n - a n - a n - n - I _ i an - ’ m u i s n- _ i da t i s L axu r m m i le l mt a fmi l e ’n- in t i n- u ts- c-n - I n- Liii and A l-I - n - ND n -u rm n-c ’ m m ‘r ’I m e mt n -u n -j i m u Ill t’aivi 0cm-n - in pi’ivi Ls to t r a ve l tb n -em 0 11- n - m s - CIJAN CE a mm - 0 n- Ni i LJ i ’ to As sign - - ‘ e” m i i-n - u u n - n - - an ’ at u Ciie’” m l tme’ c-ssar’-’ 5 er n- C n-n-- o r ’- ’ HI _ x uctur ’’ ‘ ‘i V a c y Dii The d et c a l a m e - Lm n -m’ - L N 1 tOS 10 4 2 m n - m t s — L a n ’ m m e ” - - m I _ L i e n - a n - m n - i -s i c - a f u- - i l i r a s s - mO ’ S c ’ ion r ue -a n - m ’ s ’ u l au ’ d i n u - u t e u t ime - na- mien - u n mi c-mitt A U s e r ‘ S n -n -u n-’ s t - m i c e u n-er iii ’ ia-s sub—cl t au r e c t o r ie s of arl - m m m i in Fig 10 7 - may m e -p th l’ m q 10 7 — D ir e ct ory u r th er i i o t ii’ a a I iv chin-in-led into mis is ii i u s t r at e - n - I ’t SITE I - T I I- be fm- - s i n - i l a - files i s en c a t l i tiSEI- lCl i- R SI m ’ m t l 3-p LREC’ j s iC i- ’r r i St 1- ’ l I E r ’t i s m t u i ’s ’ m m ‘ t’ — I III I’ m I for l- ’ acn- ’- inm i Ic Iii — — - -- --4 - — n - m n - ’ m u- n-c-’ ma - iri s for in tim e U u rc-c tom ’y t-c’c m fy inn - al l of am mriedu- ’- r inn- p o i n t ma m m i ci is in irs - There m n a i n-i a- t n - e n -n- n o d e ’s The 01 n- m rt inc j at Li m e directory -m’ sI i i sed can -ms is ts n a me t h e a n c m ue- r tree t i m e ’ ‘Pop Co n t e x t ti me n - t I _ i c - i a Context ‘ u ’ o i c oi a t e ’x t a n c h o r s t i m e p a t h n n -r ce is use-cl primarily f o r m’ e l c - r c n c i n q t i m e ‘Plic m c ii ‘ontex t Att system aut - mm tima ’t- u cs im i l c m e c ha n 1 srm ‘l Ime ia n - cL ’ n c-u i ’’ -in n - ii- - i s - ume-ne r un - n - me re t c - ru -fl oe’s n-ne a m a t h ‘im t ‘m a t u ’ t c - c i m m u j l e user r’ oaths via mm fu r - n - - ’ ’ c - n t s - x t m u Css i mu e -c t m e- ut ti m b e ’ iccessed an -In - l ur muucli t i m a - d c - s i r k -n- 1m m - I — re-ct he ss ’ l u a - n n - - - ‘ t t 158 ut m ttc aily is set E n-’ is a p a t h w h i c h — the b eqinniny of a session T he C o n n e c t c o n t e x t is mu p a t h n -nd en d s u - a - i i th a u s e r n o d e ccli is initially set am the At t m ch co n t e x t a n d t a r i -m inat e s in a s u b — d i r e c t or’ - D a t i e - u m m m i l t ‘ri m e- r amu fer n - I - mt - i las t S n -n -up ’ wi m e - s t e a r u U’ ’ d i ta t r i n c m m - I n - S I n - mi se t I i n - st u an d s at s -con cl n- m t i t -h ru — ‘-imt ‘ ‘II’ • mm ’ - - miss - in- ma m imp I s ‘ I lii’ n - h- hit -u l u m DC ha’ mm i ii t ‘me of u- uRP im m u- ‘ ‘ u -m I ‘AId - u t - - mm in Is ’ n-n- I an-ill c a l l n o d e f - I lu - a- c-aI use-i’ All duta anal retr rc- ’c - u t im - m t n-nt Cs time -- i t ‘ this control strc’ai i has time to a-mm tran smission transfer tm mm 2ii2 S Ire of l0 S Ic FAXDO C rn - -t n - m en - cr 1-l i - is alsLn q TIP ml In I and SEC RET f l u ‘c ci en- to -m ccc-sS t Ia a’ it u s u n - n - m i m e - u i t h e I n - I A n a n - c - m i iil e- e- n - n - u mu them ul- itn - An example is n - t e e n in Fj mc mmmimm - -l to When p it h the f i t s ’ a t i ’ m m c t m m i ’ u- c-I in- 10 7 i n - a c i c - u r l y d c - l in e d i m - m l n - - mm a- w t t m u des ij mn --l In-i u - c a m I u mt s - n - m ’ u thei m a n hmur mmu n II ’ ian -- - s -r l ’’ 1 1 u — L-im ’ u - h Im -u- c’ ron-ic p a r a nm e t e r s un -ic -n ea l be’ CS In - u lu- mm l - Il-nn - ’I’l - I ‘Sn e -u ’ ic 1 - o r t i o n ur n - cnn n-s 1 mm- i II ia n - c - 5 1 c - uI n - s- - i n j u l u u r in - - u n - I r e 11 rm -f C l’ us E S t l s ’ u mt i ‘u i - l i e s I n - h n - fin - - mn cI - S u u n s mi s s i on u pru - i - n - re- - m r FIIUS’Pl- l N St Protocol” tin-cd to control is St mc - iii ” n- m - t h i s d- uIn - u a i m C a n i i c ’is tm ion 1 u m l I I ui n -h em i n-m i ’j ’ -ui ucl n - m c ’ n - m m m n - - c I 01 — mt Da tmuc u -r m cu ut ei ’ ti m e c c - m i n e - c t to OflO tin --a’ lii “ ma1 n - i u ’ a - r m l u i u t e r S Tim is n -u n -mne cti on n-lust be acti -- c - t h r o u g ho u t t h e “C e nt ro l n- i ti c - in - e’ xn - e ’rI’i - l l im rm ed - ‘ m m- -m m - l y e - mm a p i - r o n — lute ’ time S t r L- ufi t - s t mimi- u u c t n- ma e - o r np u n - e i mc uS’S is ’s ’ m m pr oce ss tim e it in - Ii f ol 1n- ii c- d U’ a l m U N m oon 1 - Use of ti n - - 10 4 3 I-I - 10 ’ I _ - - si m w u t I s poss C i In- i t - t o - ii- i m c - the - -- - ‘ “ ‘l’Ii far tm ’u e n-c ft n- mm a n- in - mn - I o n e e m-m n - n u n LO n -lI m -n - - m n - - sm “ is c-f c_c n- I - - t a U se -cr ‘ ml Cc-1 u al COn - m u i C t c ‘u a a -m tEe s san -re ru n- f o r m a t anal cm ea - cel ia-nt ‘ -- -n- i m t o l l s a m e i i t - @ I 0 a sit e t Ime ‘ u I c- m i ’ -f ‘l’hey -java - tm’ n m m m mn - i - ‘ mmmi jus t the- r - -h m 4 s l ’A -m L W o i d ” h r t a ’ J IlL n-C ia SECRET - -‘ _ _ While the - PORT i n f o r m at i o n cn- t o soobet Datsic’’na1 u t e r C -in imec t of L o uj on to Dat mr c-r nut c-r - m n d connect t si n - n - Y E I N I m m - a l e pass te- KlRS ’iEl n a c e i c n-nd t -rcpare to retrie-vam n - n - N E c - i’ R 7 Connec t file to TIP d a t a socket C lose File Di sconnect dat _ cu socket OR Disconnect from Data— ICm ’1 i- i n - 10 8 f o r n’ n - - n - h i m c - s- r m ’or -n-c- rm m liti o ’a - - I - ‘ I’ I - -n- 10 9 i’ ç J n-tn -son -’ D I h e ’ON L m --—- t im e ” c’ ’n - Iu m muai 2 1 2 1 I m t ’L mm R I- n- u e m u- ‘I’ u in Fig n - Ln - U5 m a n - c d U C t a l r c - - 1 97 1 C CA n - u t of f a n - u n - en- in- D i n - - -L m n - mua c c r emm j on mn -e u S _ i i ’ s _ c mn - m i S _ mic a t h i rd ‘i an - a - s lu - a m i m re a immn le CONF- i- - ’-l’ I- ’AXI’ok’i F’A DOC in A - r b u r n - a n s this- v are re r ’ u -fin -mac-I m m c v ha mc t la’ it ies a- ’ I tc- ciusu thcr n-cnn -ut c- i or m n -n- mm ’ hm ’ - s mm si’ L E n- F L n -a ’tS 1 0 5 is shown l - u n u c m- -- m cm S ‘3b CR n- I- T N -1 C 1a u_ -mn -l ’ J c m r v c a r b u ss_c L- c- IN 11 d I n - H A CL O SI Eat t m em’ m ’t’ _ — 31 -2 R 1- ’t — f a n - ut ‘ su n-cessf 1 - s a m ” - T e ” t i - n n- H to n-n-’ act s m’ ated v er y n - n - m n - i ’ d - i c - m m - mm - c ’ - 1 1- scm E n - t n - — L a - - c - a - m — p ort n - u • f l tm n - u b e i i a i o — c a ’- iu s t t n - a t t i S u mn - I l st mm m i im t n - a- ‘uc u iceci - sr I _ n - -n -- re hc- m n S Es_ c c - i n - a n - n - n - en - n tn -c tmwo v irtun - 1 i’-ircujt --u minal ports m ” 1l m c — C t C - r mO un - 5 tm o - ‘I” p r o c e s s inn- ic -i j ’j m— ham Sc- loon - s n - a - ‘ _cu n- Il a’u ’ - ’u ’n -rd f o r a n - m u i i i n- -m ars n - i a - -c-rmsu a s e S _ t a n - m l m ’ m J cm- fcm_ c f m - i ’ C n- e r s Sm sl I t ’ - -c-c t ution Scot as l l c L e the -n- ru n-c- ’ t m e ’ aI cmi i O - i a-n- is the m jm- i p t in - f l t 1 n n - c o n - n - i t o e ic r Inn-S - - i n - - ter OilCt m i m i - It s te-Lu miim - min - n- -m u i mi -r n t J t i i u u t fr ‘ Ic ’ ’ stm n -m m m E n - n - m- bn - mu- m ‘i ’i ac- tw i n- ‘n -n m - ‘ m - ’c-’J c - - ‘‘-ue m - - m ’ e - - mcd n-’- - On tn - u_c I i i ‘c- - i’ i an - a c o nu p u t i -u t i —i - - --__ -_- _ __ i u ’j t i - i - e _ _ _ _ ____ Cor an d s _ to Re-tr i c - v u - —- n-mci L i le —‘ - _-_ _______________ _ _ ‘ -‘ -9 H 31 - -CR ’ R F S - 3 i 1- CR Ci ’ nmmau - CR 1012 ’n- t it Sm J u n - e -t t n - a f in --c t i m e ’ n - i t in - u t n - t ’u ’ Trying Op es im D a t a—L ’in quan-sn- r c c - a - n c - e ’ ” - l’ C’i ’i- I 003 1 770 1101 3 -4 2 5 J1 jO 770 110 13 5 21s 00 41 770 13013 502- i n - n - I NN D A n - ’n - u ’ m l ’ i l i l ’ I T J 2 0 0 7 7 0 1 1 0 13 5 0 2 5 i RIIRI JN k i n - - n -I ’m ’ 1 2 1 0 77 01101 35 l2tu LAGC ki AD1 N C LOG iN F A CE I N I L E 1 - i L ci— L anmj U ci i e’ 0032 -1002S 770 110 1 -10028 1200 770110140028 7701101 - 1 0 02 8 • 120 1 ‘i- 1 ‘ P l o - m n - - -ES _i ’ 770 1 10 1- 1 0028 J2 Y 0 i I ‘ I ’ Ii - r a’1ni mi m lIn - r i I C m - t t i _ I _ _ I ’S 77 0 u’ -n- ”’ 2 1- — - n - i - ‘i’ ’ Nm ml’i’m ’ Y uJ n -m iNi ‘ s oWN -d ’ l ’ ’ ’ hS ’I’ 1”OR Lil a i-I ‘ CCA It-i 2f’-DS tn- n- El i In - J Cl-t 1 1 2 a um ltR IN IIOST ’ L O N I X N — ’F’LP A n-LOG IISER ’I’A CSU-lIhi ’ Rh —tI N ENL -I CL ’F iON RIIRL I N RE ADY I•oq in s-i I r n j I s -imn ’ nt i s d i i s - v e - _ 10 ‘ui ’ En SOCi-1 1200000 CO i S P h l ’ t ’ b FOR REQUEST I’ r o r ’du r - i - - mr ml pi’em - u s -n t _ s ’d - i n Iu iV e Se ‘ REAL IN t Nin - - ’ DL II I 1’1-’ER r AGc h i t i s s m m s u t ‘I - li ’ - Vs ’ Om ’ _ I Q 1 1 3 ’ I-Oft L I A n - I ‘A RY ‘ ‘0 - ram apon cme ’ 7 7 0 1 10 1 0033 I N L ’ I ’ I ct Ion u i V s ’ F V I e ’ -•’ cml 10 3 we En-vs_ c out l ineal -- ‘ - - -- -- what imc u ’u —— ‘ ‘u el’ a r c’ the u n - a r - _-- __________ - w ‘ - - - ‘m n- ’ I ’ ta ‘ ‘ ravn--sm ’m ’ It c-s e’ln -- -’iri ’ m- m m ’ m i Sc In - am - n-nd It elm a im am c- Ii 5 I cie- m cr mn - n - - hr i c i L y J I u ’i n- I S _ i c t ’’c -_ mm ’s_c i 10 3 2 S_ l i n- 5 lu I Vt’ t _c ‘ - ’ im p ’ it mm s t il e - ‘ ‘n- se - - S _ h‘ th e m - -c ’ 1 2 on-t - ’ - -c c - ‘ - c t n- h “I I - ‘s’ - I - - ii 1 -i i k- m I in - - m I m 1 ’ e— - - Iliac ‘ ‘ a- -i 1 1 i s L - - t-’s ’sSSt le i - C TIP of mime a a no -an -- ma o ’ _ ce n - n I- C- ’ I i lt i c - I i i hic i’t - t 1c’ ’ c- -olin- - -m” I _ r _ c L em’- a is - a s ‘n c-mn -a-r - - ‘ ii - u o’ ‘ ‘ 7m - ‘u m S n - ci e_c e ’ t i ’u c ’ ’i in -- ’ ‘ac— Sa- ’n - i on ’’ c’ I ’ LL ’ - - n I ’ ’ ‘a t re - s ’ c m f l b-_ c ‘I’ l u - c - I m - ’ cmc ftw ire 10 ma - t im - _ c Sn - - n- c i t a u-c a --i c - en - end 11 ‘u m m - ’ ’_ i - i f i c E n - m a -i t ti s rt pit r’’e m 1 ’ m n - t h fo cor u i it n - ons -i iss t cia I I’ m P Q mi - i’ T - X n-c be ’ ‘ n - l t i i i l ’ - a’a i ‘a’ - ” r i t e-aJ In el c-S 1 l 0-- u b — cmiv i - ir a - ‘S_la ’ - here -i - n - ‘n- n - i 1 1 c ‘1’ - ‘ ‘ mm m c ’ m i u5 a1 m n - m o 8 ba sic c lcn -mem a t 1 Va’ ii ” FLo - 1 - 1 - di I-m i i ndlm -i’ 0- i’m I’ i c ’ i_ I Den-sm’ - - m n r ivers C a F m ’ - o f i i c m m t i - -i Cer - m tr - -- lIe r ri te-c f u c ’n-_c n - cL — i - -m n ‘ 8 - 1 J n - m m t -m u ‘I’ -1 7 il mc ac ‘u n - ’ ec -r l v s u ‘ sI c’ - mn - _ ct i n- i- h-tn - mn - n - a - I t ui -m u c1em t c i l l m 0 mn- i a- -‘ 3 c ’ ‘ 2- n i n - o l a ’ I’ lc c ’s_ci S 5 ’ - - ‘c-m n-id n- 1m m - I’ im rj_ u len -’e rmt ca t icmm c-F L Sec t IOn-i - ‘s a t m n - - i c ’ -I- - 10 iO - - - the c u o n e m i 0 - - ‘n-i e-fl i rm’ e r i t s - n’ mi - S o f t ’ u m r - 1 n c_ m r t - i F i c I-m -i -m m In m n a ’h’ ’ m ’ nm s - i’ • ‘ I - - co ’m ’ m - - i 1 - t l smm - ‘ ‘ I a ’ n- mu — t r - n-u-i ci - t h c - s ic -- i in-cs 1 1 1 1 L - n’ a - i -va cm cj m lm ’ For - - not b - e e c c - ma in -n-s - - c ur - it i -ia mm - c-a ’ - - 10 4 3 I _u -” ’ t ’ a’V i n- ‘a e ’ i - - nc _ c ‘mm _n - o n-a - mit cal m-i n - m i iS - ‘ ‘ l - c eti’jevai ‘aith _ cs - ’”o -t’iaLC n -ma- - Ia g iv e -- i 1u 1 - icr ly m tructure si_c ’ u a i J i i I __ cO t -d 1 mct ma m i 1 u ard ’ -i -mmm i’e m e ’mu f i - aur-rutiors n-nd t h e 2 t i l e - ‘ - 10 4 3 mm t i e s i i ii ‘-- i t iO mi n - m n - i - i - ’- s i u i mu 11 1cm -a ’ 10 ‘3 1 -n-mi d ‘ e s n - a - m be - an - mm n - -’ SO c -r i -n - U n-i - - ou r i n - -n - - a - n - s Ii- ’ o u — N c m c i r m u l i r e s Sj o f l n- t n - e t -— i i p1 011- c - n - s tim e t oe - sim ile pr oj - -ct most of the t ime was - - - - - 162 - — - - — - - — - ---------—-- ----- -- ‘ I i - -a---- H1 - - I —I cm0 1 I - - C 1-i I- - L - c-c In c—c - a - a ‘ - - - n-mm —lam _ __ _ j _ 7” _ -‘ - T’ ’ _ ’ —-- -’ - ” ‘ L 1 L _ _ _ _ - IIII L L _ - - _ _ _ _ _ _ _ _ _ _ _ _ _ ‘ L _ _ _ _ _ _ _ — _ _ - ‘ — ‘16 — - spent in u’ 1 t’v s ’l op mac t h e abuv ‘ - -i a m - fl-- -flt s to handle sta n -in -u to disk an-u ’ a de n - i c - c - t o operate in a aicU d aIX in - si c l iii a oe - ccsuse of time neec ov a- u in-p p r a - - m’m ’n - m e s s a m g n i n i cn - n t n umber m - f and tasks can T h u s there in- ma-n - i fm i t si sin -sill r e a l — t i m e e x e c u t i v e 1 whose main fun -u ct i-ams are a schedule the a p e r a t i o n of the software be c o n c u r r e n t mo du l e s act a s controller and rm ’m aintain an efficient traffic mm d a t a f l o w b e tw e e n different m -roce sses The executive provides b l o c k s on w h i c h t h e sc-i f t uici i - ’ can be d e v e l o p e d t h e b a s i c b u t l ciiiucj cc - c- rc--luced to a small System activities n am r h e r are activated on request to t h e e x e c u t i v e is purely a sy n c l m n ’ -n - nc - mms w i t h i l l t a k e p iac e a t a n y s u s p e n s i o n can 2 Process scheduling p r o c e s s e s b eing written so that priorit v process Modules of tasks which t ime jim fmivour of a higher 5 sire standard o n e s w h i c h r e q u i r e — no c o m m e n t The Communicati ons con S_ ran -t ier 6 amons ists of the rmodules wh ich pass data in two stre mma via time II ’ P0 t o ARPANET This is the seqment data thr ou cmh The the I n -mn - erf a ca 7 m s a primitive Net - -n-n-rh Access N c t ’ n-- - c - k Machine of - ‘u iim ic had t - be cLan -n -ed fo I h e a t t e m p t to pass t h e T I P char m - m t e n i -si r t s NAN functions i re t h e o p e n i n g mind c l o s i n g c o n n e c t i o n s an a l interfacing ‘ ‘Thh the remote I t s r’ m i i n n e t wo r k subsystems e g ti n -C it qenerates t l m e m given function In the case of t im e Dat - m’m- m a - m -mmter n e c e s s a r y Data—L an - ru_ ca n - a- cm a maimnan ds ‘m u Ti-m is rmao dn-le is complex a n d access perform a is one that needs functions are to be expanded modificat _ con isis the ma’m-’stc-r The C o m p r e ss i o n — D e c o m p r e s s i o n p r o c e s s t a k e s t h e raw d a t a com ing fru m - m t h e f n - c i mma ile of r u n l e n gt h ema c n - si ing s c a n n e r and c omma ar cm s s e s K i r s t e i n l’i 76A it by a form I t also formats the d a t a i n t o a c a n o n i c a l f o r m so that a fa c s i m i l e d ev i c e with d i f f e r e n t c h a r a c te r i s t i c s c o u l d output the data In the reverse direction this module operates in the reverse n-m an -ncr 10 5 3 The P r e s e n t S t a t u s Most of the end of tI m e m i ser 197 1 fu n c t ion s of S ec -t i on 1 0 3 Wa’Uc- poss ib le at TIm e c o n n e c t i o n to t h e network was still in a rather awkward form as mentioned in Section 1 0 5 2 but this _______________________ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ — - - ‘— — - - - ‘ pp— w — --— — --- - — ‘ - ‘ - - - ‘ -i L L I be N ot r e om mae i u all cal oa r I v ui t h e f a c i I i t m c - mm ot cn - r inc im - al sh o r t c i m mm m i m m s 1 2 I ‘ 77 Te x t not in - ’ q 10 5 au” c ‘m ’c -t m c - c ’ lv n - in - c 1 1 iL 1 a’ Tn - c ’ - ta i I c- it ion There 4 ‘rhc- re is o n l y a-n a ’ I is iii the th-ssaqe - I- i l a I-Field the Subject 3 i ’ ma j -m in - tile t n -n - - l- ’ iies ar -_ c ra-strici a d P a t h Nai m m s ’ not i II u m i S c - mm - a l i t — i s no u s e r — s u p p i t e d ci -m mm p m i i m a on t h e I n - t i r o m m i - c u t o r m o n-n-inc I- ’ A Y h I M I t hat LE t m a ’r a ’ n - u - a l e’ is mmmc i tea r -ill ii- - n n i v u e n f i Ic-n - i’ m -_cram n -h e _ cr users 5 N tacm-mn - in -a ’ u ’oi m m p s i n - i t j ’ ’sl lluu e’-j c ’r i ll f- mn - elamne -n t - m l ha ’ b immi- Cc mfli’ m uiids ‘n-n-- n- a c-am ¶ iii- c - is- n - i - -rut cif n-’n- I mm e a -n - c ‘n-a m t - - c - c - o r ‘m mt i- ’m an - reason cii a ’ k nown w in - ’ c - h ’ - “ uiim -- le m ’ ’ a t or’ i r’n - lm ’r mm c -n te d iS p1 in - nm - I _ _ _ _ _ _ _ n um n - n a’n _ _ _ n - ti a - iuld _ _ nut _ _ _ _ _ _ — ‘ CHAPTER - 6 5 - 11 CONCLUSIONS It would be nice to be able to present a n umber of conclus ions and pinpoint accurately whence in this report the conclus ions have been drawn In most cases however this is not possible The conclusions follow from the integration of the different activities We ind icate below however in wh i ch chap ters research is described wh ich l eads to the conclus i ons 1 With complex software for switching data measurements are essen tial to detect inefficiencies in throughput and o p e r a t i o n in o r d e r to i n d i c a t e a r e a s f o r improve- ment 2 2 3 8 W i t h comp lex p r o t o c o l s f o r d a t a t r a n s m i s s i o n i t is essen tia l to comb ine t h e o r e ti cal an a l y s is s im u l a t ion and measurement of performance Unanticipated interactions between d i f f e r e n t levels of p r o t o c o l can degrade performance An examp le is t h e dup l i ca ti on of sequence con tr ol betwee n lower leve l ARPAN LT pro tocols and the TCP 3 5 7 8 In orde r to measure systems per formance as above a - u i i d e v a r i e t y of m e a s u r e m e n t tools are required These i n c l u d e t r a f f i c g e n e r a t o r s and l i n e leve l m e a s u r e m e n t s Once t h e t o o l s are deve loped they can be app l ied to many d i f f e r e n t networks or interconnections with relative ly little modification 4 of networks 2—8 The connect ion of serve r hosts to networks poses many problem s The UCL approach has been to attach hosts by a front -end p rocessor in wh i ch v i r t u a l calls are matched from a netw ork to an access scheme supported by the mainframe Not on ly h as this technique been shown to work well w i th rr- any d i f f e r e n t computers but a general approach to this technique SWITCH has been developed during 1976 it has been applied successfully to several host systems - -‘- Our approach would work well to for n e t w i r k n e t w o r k c o n n e c t i o n and can be e x t e n d e d to s u p p o r t an X25 interface 5 For useful 2 3 m o n - I s o flc t ’ ” n- n- n-’t attachment o f c - i ’S must mapp ings be alone at call leve l but also at highe r levels such as th e interactive terminal an al bulk transfe r support levels In t he f i r s t t h e r e are sometime s character set mapping difficulties in t h e latter there - ‘ir a ’ s o m e t i m e s r e a l d i s c r e p a n c i e s between the underlying assumptions of the net”crk s and the serve r host systems These m a p p i n g s h av e been done r e a s o n a b l y s u c c e s s f u l ly in the U CL p r o j e c t s b u t t h e r e a r c s o m e t i m e s som e aw k w a r d n es s c - s in use 2 3 9 6 Interworking between different networks poses a n n -n v prob l ems a n d v a r i o u s s o l u t i o n s h a v e been p r o p o s e d UCL believe that mapping at the virtual call 1CV a ’l is a ve ry p r o m i s i n g a p p r o a c h f o r i n t e r c o n n e c t i o n The SWITCH s y s t em is f u l l y a p p l i c a b l e to t h i s p r o b l e m The a p p r o a c h is f u l l y c o m pat i b l e wi th u s i n g X 2 5 w i t h some m o d i f i c a t i o n s f o r the 4 5 7 7 connection of n e t w o rh s computers t m r - i-m h mo re t h a n or the effective use of one network there m a y need t o be mapp i ng of t e r m i n a l and b u l k t r a n s f e r f a c i l i t i e s SWITCH provides the mechanism for s u c h m a p p i n g a n d i t h a s been in ce r t a i n 8 ‘l ’he- r s ’ is a f-- r p r-i i n - a l L i i e ’m m v i ri inmrments St ’- i in - - _ c i s c - - m n of i n t e r a c t ii i — T h e r e may well bc ’n- n - t n - s i t t ’ m b e t w e e n t h e t e c h n i ques u s e d services through en v i r o nn i en t s d a ma -1 6 c c - su e it ‘ im Os -s i networ ks f o r w h i c h t h e n e t w o r k s sire d e s i g n ed n- real dive ruencc- here between the requirements of PTT networks and those of ones mis I n n - techni ques such is packet radio military and r e q u ir i n c n u m b e r s of g at e w a y s f o r h i gh a v a i l a b i l i t y 4 5 6 7 -- ‘ - — 9 The techniques discussed in this project may encom pass a b r o a d e r are a of commun icat ion t r a f f ic tha n or igin a l l y env isage d Ex amp les are digitised speech message traffic and f a c s imi le 4 7 10 10 We have been a m o n g s t t h e f i r s t g r o u p s to h a v e used the BBN KNET d e b u g g e r to do r e a l — t i m e p r o g r a m m o n i t o r i n g ari d d e b u g g i n g v i a a tenuous link s o f t w a r e in tim e n-CL PSP G a t e w a y s C h a p t e r 7 The is controlled on-line from US Tenex sites via ARPANET for debugging purposes This approach to pro g ram d e velop m en t ha s been found to work w e l l 7 11 The development of mappings at c a l l leve l and h igh e r between external and internal conventions as in SWITCH ha s c o r o l l a rn - - s i n- o t h e r conventions need n i_c t ‘the in t e r n a l be restri cted to one computer b ut can app l y a cross seve r al compu ters some of which may even be remote This erodes the dividing l ine b etween a “ s in g le ga t e w a y ” or a ‘ g a t e w a y system” This ‘ g a t e w a y s y s t e m ’ may i t s e l f be a compu ter ne twor k “ single computer ” The whole pr ocess b etween “ computer networks -mn - “distr m i uted computing ” is hereby becoming blurred 2 3 4 12 In seve r —a l projects at UCL h i g h leve l languages have been used on small computers via cross—comp il a t ion Babbage on the PDP9 RL 360 and BCPL on the PDP11 There is s ti ll no clear ind i ca ti on t h a t the approach i s r e a l l y s u f f i c i e n t l y convenient in operati on or e f f i c ien t in code produced to be really v iable Even in our own group the opinions still differ Some co m p la in that the awkwardness of having to use several s y s t e m s together c r o s s - c o m p i l a t i o n there the code produced factor that manual outweigh _ _ in is a l s o clear evidence that is larger and slower by a s u f f i c i e n t reco ding may be necessary and too much extra memory required _ the advantages - At the least a high— - - ‘ -- _ _ _ _ - c - a-m n - Lai n- c e n t ri l in -n --ic The _ _ - -‘ 1 ai- u c - l i l i l i t n - ’ -‘ iule -“ - - ‘ a nd mo st ue - 1 ’ un- m ia ’ i n - l s -n b- - mi sOd - n - m i code is b e i n o produced or t h at m a c h i n e m u s t a l s o be o n — l i n e Tl’e p r i n c i p l e UCL - - a an - n - em’ cn-c- tn - a_ c i n - t n - n - n - s t m c c o m pat i b l e w i t h t h e m a c h i n e f o r w h i c h 13 — — ‘- -‘- - - 2 5 7 8 10 a c t i v i t es in t h e P a c k e t S a t e l l i t e P r o j e c t have bc-c-ri to develop t o o l s f o r t r a f f i c g e n e r a t i o n ano M e a s u r e m e n t t o o l s have -em _ c developed b u t in the absence of G a t e w a y c o m p u t e r s at o t h e r s i t e s d a t a a c q u i s i t i o n no r e a l UCL m e a s u r e m e n t s have been made in 14 1976 7 The user leve l measurements are g i v i n g s i g n i f i c a n t i n f o r m a t i o n on p i L L - i ’ i ’ i of network usage on the usage characteristics of different app lications and of the ps_ crforrnance of all the subsystems available The measurement t o o l s s h o u l d be run consistently and continually and p ref er a b ly cove r a l l types of access for highest credibility 15 8 9 Comp lex traffic generators data acquisition and measurement d a t a r e d u c t i o n a l l r e q u i r e s iz e a b l e a m o u n t s of ware of computer soft- C a r e f u l a t t e n t i o n to f l e x i b i l i t y in t n - c d e s i g n s t a g e the se tools p a y s o f f i n the eas e w ith w h i c h t h e y ca n be extended to new un anticipated app lications 2 5 7 8 For efficient network usage facilities such as message 16 sJstems and file transfer are essential it is e s s e n t i a l also to provide adequn -Lea user information on status of connection and jobs and on use of facilities 17 9 Usage of a network for collaborative work depends heavily on the availability for each of the collaborating q rc-ups involved of a local attached host 9 T h e facsimile services envisaged would be of great 18 b e n e f i t to ARPANET of course in w i d e r the f a c i l i t i es u s e r s- t h e i r p o t e n t i a l n-re f a r g r e a t e r contexts They could be now a v a i l a b l e on A R P A N E T Initi a l economic s t u d ies look promising for private networks development on PTT networks srovided with — wider is c l o s e l y bound w i t h PT ’l’ tariff a n d policy consideration 10 ‘— — - ‘- -- _ ___J____ _ _ _ ___ ___ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ _ _ -“--——--—- - ---— - ‘ -‘ - — 19 Howeve r w e l l i n f o r m a t i o n status facilities and ba sic ne twork se rv ices a re prov ided it is e s s e n t i a l that personal user support be available Clearly no new serv ice c o u l d be in t r oduced w itho ut cons id e r ab le a t t e n t i o n to documentation on user facilities Howeve r our exper ience is t h a t almost eve ry user in c l u d in g these ins ide our g roup comes to requ i re personal intervention of our user liaison services at some time The amount of effort required to solve the p r o b l e m s is v a r i a b l e i t may be a s m a l l piece of i n f o r m a t i o n it may require a few messages to other s ites it may require a large UCL software development or it may be insoluble and suggestions to circumvent the p r o b l e m may be n e e d e d Ho eve r we b e l i e v e t h a t without a specialised liaison officer knowing something about J J PANET UCL - and RL etc most applications would make significantly slower progress probably the majority would give up deeming their problems insoluble 2 3 5 7 9 20 Given t b ’in the L L leve l of a PA N E T work i s now b e i n g t e c h n i c a l and user support provided p r o jec t ve ry s ig n i f i can t and u s e f u l j r— n c c ol l a b o r a t iv e l y Some of t h i s work w i l l n -e t r a n s f e r a b l e to c o m m e r c i a l s e r v i c es by 1978 o t her s w i l l n o t b e c a u s e of t h e unavailability of the r e l e v an t US sitc-s via international commercial services both f o r economic f a c t o r s c a u s e d by low average usage and by local p o l i t i c - i l con s t ra in t s 9 fl ‘ - i 7 0 RI - c - m ‘ - - s n - L t Pi-ava i c a l R e v i e w L et t er s 37 4 AN DE RSON EL 1 7 O at e av b c-si gn f o r C o m p u t e r N e t w o r k B E E L E R M et al I 75 - l t n - i - L i s I n -t ‘ In t e r c on n s -a L i on H I - b Mass E s t d 1 ishment of n -he B l a c k n e s t n - t h e r l o r d L a b o r a to r y ’ s S e i s m of o c L c s L b a n - sh a s e a on _ Bhi ‘kin -- st bes urch E s t ab l i s h m e n t 3 6 0 1 9 5 CL m apn - ti cL BLAME Y C lt 7 m ‘rha ’ - Brimpton n -- R c a _ c I i n 1 ick -t B ROOMFIELD CT 1975 SwLL ai juu °‘mO i - m s — Exnearimental Packet ka -v i ew S c-t h _ ca Se a ’v mc e C c - ’ o i n 1 ‘ - r— - -cP A o iO- P fer tn - BUTTERV ELD S 1 97- Netwo rk u t Proc 7th iJaw iii I n t i’on’ S t’ ’ Sciences 1974 - D a t u c -’ ’ O - r P r o j e c t Work L fl u V e r n - i ri P u -’ - r CDCA CDCB 1 9 76 ‘- t a n i l i -mamp ’ _ ctE r -Corp - Cambridge M a ss CCITT s- c r I ’ 19 7 6 - CCA X25 0u t a c o m p u t e r America Rcc -raumic-nU u t i o n AP Vi h o 5 5 L Geneva Gen- m ’ ui ln fornm u t i - rm on tim e n-se o l I N T E R C O M ULCC 1 976 — B u l l e t i n B2 5 l 19 72 T m a - C o n t r- i l 0-itul 200 User Ta-i n -mm mi -i l P r o g r a m m e r s Guide docunme -’m t ‘ - i- ’72 ULC C C E R F V a n d KA Hn - R A i rotocol for P-acr et Network Intercommunica ions l’hE ’n- i’ ins a ’ O r _ c n i u i l i c a t i o n s - Vol CO M —22 N o 5 p 637 CERF V 1976 CERF V 1 9 7 7 DRINKWATER B l 74 Report in - prep arat ion A n - PA Arlington Vir qinia Report in preparation ARPA —rli ngton Virginia Use of l 97 1i Building Description P o ly t e c I n m c EDG E SW 1 977 of h c-r - n -n e t t em s D e p t of of Op t i m i s t i c ann -i Conservative Credit Return h o s t — h ost Protocols with UC L 1 97 Several Netw- rks with a Gat -‘ - - m r m u n i c a t i o n m m HIGGINS iN PL and FIINCHI LY Ad In d ra Note ‘ A s’u u y 4 1 4 D ep - ir m n - - c - m -t The problems of linking European Conference Ce - rn - ou t e r N e t w o r k s n-K p 4 53 19 7 5 B UC U - Resecact to Flow Contro l Del’artm ent of Stati stics and T i - r a c - a u t e r Science HIGGINSON PL and HL-A ’iihi- Y AJ on C o m p u t o r Mathematics o f l cl O f l Evaluat ion Indra More 5 9 4 A n a l y s i s and for Survey ot S WI T C H — An Overview Statistics and Computer Science - HIGGINSON PL 1976A CEPT HIGGINSON - PL ESP 32 19 7 6 B C o m m i s si o n s R e s t r uc t u r i n g of ESP 25 EPSS a nd X2 5 - A Personal View Indra Note 569 Dept of St -i tistic s and Computer Science UCL HIGGINSON PL 1976C First Specification for SWITCH FTP Indra Note 580 Dept of Statistics and Computer Science UCL HIGGINSON PL 1976D A s y n c h r o n o u s T e r m i n a l s on E u r o n e t S t a t i s t i c s a n d Computer Science In d r a N o t e 590 Dept of UC L HIGGINSON PL and IIIN CH L E Y AJ 1976E Plans Dept of S tat i s t i c s a n d C o m p u t e r S c i e n c e PL 19 7 6 F 4 6 1 Dept of HI N CHLEY AJ 1 97 6A Seg r -mn -’ nt L o a d i nc on Science UCL AJ l976B 571 D e p t th e UCL PDP-9 Indra Note S n - n - t i n - t i c s and C o m p u t e r S c i e n c e n-CL T a s k i n g an d Se- mn - i-m n t at i o n New System Indra N o t e 4 1 4 Dept of HINCHLEY u pp o r t i n g t h e c-C l G a t e w a y I n d r a Note 531 F i l e T r a n s f e r P r o t o c o l s of HIGGINSON For SWITCH n-f S tu t i s ti c s - in t h e PDP- 9 S t a t i s t i c s an d Computer R e v i s e d S e c i f i c a t i o n I n d r a No t e ari d Corn u t e r S c i e n c e UCL yn - In- JEI N C SYSTEM RELEASE I n d r a HINCHLEY AJ 1976C’ S W Ir A i HINCELEY Ad i576D Intr _ c - iuction of Gateways into the Satellite Note 583 Dept of Statistics and Computer Science UCL Project - Issue s Arising and UCL Plans for Working with Ga tew ay s Indra Note Science UCL 537 Da p t JACKSON PM and ALCLIUNAS R 1976 of Statistics and Computer Decomposition Based Generating Functions for Sequence s Den t of Combinations and Optin isation University of Waterloo JLSSIIOPE CR 1976 A r p a n e t and the Th in-c IV RCS 53 Department of Computer Sciences Readir - University JOHNSON JO 1176 Indra ’ s PDP-9 System Performance Inorove ments I n d ra No te 5 9 5 Dept of Statistics a n d Computer Science UCL KENT S 1976 Suqgestions for a FAX— NAM Implementation I n d r a Note 5 4 4 Dept of S t - ti stic s and Computer Science CCL —-- - - - — — - - - - 1 72 KIRSTEIN PT 1974 - n-CL ARPANET Project Annual Report 1—7—73 to 30— 9—74 TR-17 Dept of Statistics and Computer Sc ience CCL KIRSTEIN PT 1976A CCL ARPANET Project Annual Report 1— 10—74 to 31—12-75 TR—27 Dep t of St ati st ics and Comp ut er Science UCL Gateway Experiments in Connection with the KIRSTEIN PT 1976B Packet Satellite Project I n d r a N o te 574 Department of Statistics and Computer Science CCL X25 Based Experiments in the Context of the KIRSTEIN PT 1976C Packet Satellite Project Indr-a Note 578 LINKABIT Packet Satellite Program Plan L in k a b it Cor p 1976 San Diego California LIN-NAN LEE 1976 n - t m -m e a l Aloha Demand - - Assignment Access Communicatii ’-r Protocol with Fixed - A Multi ple TDMA and Slotted - Aloha Features Linkabit Corp San Diego Cal i f o r n ia LLOYD D 1976 A Note on the Role and Requirements of Gateway C ompu ters Indra Note 536 Dept of Statistics and Computer Sc ie n ce UCL P0 1976 PO-TG16 Appendix E A Restructuring of X25 into h-IDLC INWG Protocol POUZIN L 1976 Note 38 Reseau C y c l a d e s ROBE RT S 19 7 6 In t e r n a t io na l In ter commun icat ions of Publ ic Packet Switched Networks Proc Communications Toronto SEXTON 1976 of t h e 3rd Int Conf Computer p 239 End—to—End Protocols Virtual Calls and the IIASA Network INWG STOKES AV 1974 Protocol Note 46 A Users Guide to Babbage TR-13 D e p a r t m e n t of Statistics and ‘ -a -nmo ut cr Sc ience UCL STOKES AV and HIGGINSON PL 1975 The Problems of Connecting Hosts into ARPANET European Conf on Computer Comm Network UK p 25 STOKES A V 1 97 6A Mon it or i n g a nd Access Con t rol o f the London Node of ARPANET Proc NCC 1976 pp 539-603 STOKES AV 1976B XFSTATS A User Guide Indra Note 552 Dep t of Statistics and Computer Science CCL STOKES AV 197Th TREA DWELL S 1977 CC L N ode of the The IncIra Network ‘-loasurement Project TR— 30 Me- i siire m l Characteristics of Traffic in the ARPAN I- T I n d r a Computer Science CCL Note 582 Dept of Statistics and — 173 — WALKDEN F MCINTYRE HA LAWS GT 1976 A Users V iew of Parall el P a r a l l e l P rocessors p resented at CERN September 1976 to be published WALLEY PC 1976 Description and User ’ s M a n u a l of CALIB RA Work ing Report 76—3 Br u n e l Un ivers it y WI LLI AM S P 19 7 6 NCP Performance Measurement IN DRA NOTE 584 De p a r t m e n t o f S tat i st ics and C ompu ter Sc ience CCL ZIMME RMAN H 1975 The Cyclades End-to-End Protocol 4 th Da t a Communications Sym Quebe c City p 6 -- _ _ _ _ 74 - APPEN D IX A Applications to Use ARPANET approved for Use during 1976 N ame O rgan i sat i o n Pr o j e c t Si te Used Mr F Grover Blacknest Research Est Brimpton Read ing Seismic Data Exchange ISI Dr J Alcock Dept of Physics B 1stol University Tyn da ll Aven ue Bristol Collabo r a ti v e a nal ys is of h igh—energy scatter- LBL Dr A J Harley British Library Lending Division Boston Spa Wetherby Yorks Interlibrary loan network development NLM Dr P L Holmes British Lib Res i Dev Dept Sheraton h ouse Gt Chapel St London W 1 CAN CERLINE database evaluation NLM Dr P Humphrc-ys Decision Analysis Unit B r u n e l Un iver sit y Kingston Lane Uxbridge Middx Decision Analysis software development D r J F itc h Cambridge U Computer Lab Corn Exchange St Cambr id g e Collaborative development of LISP comp il e r f or al geb rai c s y s t e m s ISI Prof M V Wilkes Cambridge U Computer Lab Design study for a data ring ISI Dr M D C Dyne Dept or Engineering University of Cambridge Decision Analysis sof tw a r e d evelo pmen t ISI Dr G D Cain Dept of Electrical E lectron ic En g P o l y t e c h ni c of C en tr a l Londo n 115 New C a v e n d i s h S t London W 1 Di gi tal Filter Design Techniques ing data ISI ISI ‘ - ‘- -‘- - ---‘ - Name Organisation Pn -C Lu D r R Pye Communication S t u d i e s Group Ass a ’rm nma -nt o In - i t a c t i u ri between Travel Ta ’l ecomn ’mun- Dept of E n v i r o n m e n t a l Stu dies Site Used ’t — O FFICE—i ie -itiOflS UCL ALA C u lh am Lab Abingdon Oxon Dr N y col l a lo rat ion I -u s c - i n Mr V D a t a b a s e e v a h m i a t j i n N LM Collaborative dcv of high energy database LBL In- ave pm ’ m ent of proof-generat ing system SC—Al Dr J Dirlington Dept o f A I SRI—Al Edinburgh Univ ’ am ’s t v Program correctness proofs program sy n t h e s i s P rof D Machine ILL — NTS R o b e r t s Pinkerton Technology Report Centre St Mary Cray Kent Dr F D Gault Durham Particle Data Group University of Durham Mr M Gordon Computer Science Dept Edinburgh University M ichic Machine Intelli gence Res Unit Edinburgh U n i v e r s i ty Res r e s ea r c h I n t e l l ige n c e BEN at present Dr G M Bull h ead of Computer Systems Hatfield P c - l y t e c l m n i c Coh en -a_ c L a n e Hatfield Ilerts Development of BASIC Sta n d a r d s NBS Dr J B at e s Inst of N e u r o l o i m y N a t iona I h l osp i t i I Queens Sq W C I N atural language ana y s is s o f t w a r e SU— A l An a l ysis of neuro— c h e m i c a l d ata SUMEX I r D M f3owt-n - Dep t of B i o c he m i s t Inst of b a - u r o h o - y National hlosp i t - il y 1 - — - - ‘- - Project S i te Used Mr L M Popovic School of Electrical Electronic Eng Ki ngs ton Pol ytechnic Kingston—upon-Thames Surrey Program generation by forma l general me thods MIT—OMS D r T W i l k i n s o n Ap p l ied P s y c h o l o g y Unit Medical Research Council 5 Shaftesbury Road Cambr idge Neurophysiology c o l l a b o r a tion BBN MOD Nc-rthuinberland House Mr D Curry Collaboration with US Army M a t e r i a l Command HQ OFFICE—i Dr N J Raitt Stud y of Ionospher e UCSD Mr B C Rowe London P o l y C om p u t e r Unit N London P o l y t e c h n i c Holloway N 7 Su r v e y a n a l y s is software evaluation ISI Dr R P Johnson Chem istry Dept N London Pol ytechnic H ol l o w a y N 7 Program development of C A Synthes i s desig n of organ ic molecules HARV- 10 Dr T Qu i rk D e p t of N u c l e a r P h y s i c s Oxf ord U niv e r s it y Exchange of software of data f o r h i g h en ergy ph y s ics ILL—NTS HA RV— 1O Mr S J Hague N u m e r i ca l A l g o r i t h m s Group Oxford U Computing Lab Collaboration in Development of numer i cal sof tw a re library ANL Name O r g a n i sat i o n 176 W C 1 Mullard Space Science La b Dorking Su r r e y Mr G Coulouris Investigation of Dept of Computer Science d i s t r i b uted processQMC M i le En d Road E 1 ing developments RAND-UNIX C M U—A UCB P r o f R W H ockney Dept of C S c i e n c e Reading University Whiteknights Readi ng P a r a l l e l processor evaluation I l li a c IV ANL Dr P Purcell D e p t of D e s i g n R e s e a r c h R oy a l College of Ar t Kensington G o r e S W 7 Program d eva lopre nt for b u i l d i n g design UARV-10 CMU MIT-MULTICS ‘- - - - -“ -- ‘- - -‘ - _ _ _ _ _ _ _ _ _ _ _ - - Project S i t e Use D i g i t a l signal processing MIT—Al Utah M r N Neve Room L121 Royal S i g n a l s Radar Est M a l v e r n Wor c s Su p p o r t f o r DOD evaluation of CORAL I SI Dr J M Taylor Royal S i g n a l s Radar Est Christchurch Dorset N etwork security NELC MITRE RA DC D r J Cr enn eh l Ru the r f o r d La b Bubble chamber facihity data exchange LEL Prof F W a l k d e n Dept of Maths Un ivers it y of Salfor d Fl u i d dy n a m i c s program development I L LIA C - IV Mr T Crowe S y s t e m s a n a l y si s D i v Thames P o l y t e c h n i c W o o lw i c h S E 1 8 Relational database i n v e s t i g a t i o n Datacomputer Dr L Koh out Neurochemical data analysis SUMEX-AIM Development of ALGOL 68 Compile r with CMU CMU Name Mr O r g a n i sat i o n l R Whitworth Roya l Military Collegeof Science Shr ivenham Swindon W i l t s UCI-I Medical School St Pancras Hospital N 1 P ro fe ssor L M De lve s Dept of Computational and Stat Science U n i v e r s i t y of L i v e rpool Brownlow H i l l RAND -UNIX Live rpool -— -- — - _ _ _ _ _ _ _ _ — rr _ 178 _ _ - - - - - — APPE ND IX B Non—Act ive Users During 1976 N ame O rgani s a t ion Project R eason for Inactivi t y Prof M V Wi lJces Cambridge U Compu t er Lab Design s t udy for a Data R i n g Lack of time D r M D C Dyne Cambridge U D e p t of Eng ineer ing Decision a n a l y s i s s o f t w a r e development Cost of t e l e p h o n e c a l l s poor qu a l i t y telephone lines D r G D Ca i n Central London Poly Engineering Dept D igital S ignal Process ing Unable to make US contacts for Dr K V Roberts Culham La bora t o r y Plasma Physics Compu t a t ion Lack of time - and of the ava ilab il it y at Cu l h am of ARPANET host Prof D Michie Ed in b u r g h U Machine Intelligence W ork carried ou t dur ing visit to US but collaboration Machine Intelligence will resume shortly Research Un it Dr J Bates Inst of Neurology A u t o m a t e d A n a l y s i s of Natural Medica l Text Illness Dr D M Bowen I n s t of N e u r o l o g y Analysis of Neurochemical Data at Stanford Lack of time Mrs L Popovic K i n g s t o n Pol y Schoo l of E l e c t r i c a l Electronic Eng Formal and g e n e r a l methods f o r formulating technological problems for computer isation Illness Study of Ionosphere Work carried out Dr W J Raitt M u l l a r d Space Science Lab Mr I R Whitworth Royal l M i it ary C ollege of Sc ience - - - d u r i n g v i s i t to US Dig it a l S ig n a l U n a b l e to make US Processing - - -- - - - - -- - - --- - - - -- contacts for collaborat ion - - - - -- - - - - - - - - - APPENDIX C P LJBL ICATIC 4S AND PAPERS PRESENTED DURING THE REPOI ING PERIOD Developnen t in European Public Data Networks KIRSTEIN PT Reichnernetzc un l Datenfernverabeitung InformatikFachberichte No 3 ed D Flaupt and H Petersen Berhing and h eide lberg KIRSTEIN P1’ Springer—Verlag 19 7 6 39-58 Data Conriun i c-ati on Ceneral Princi p les The Encyclopaedia in-f Computer Science ed C L r4ck New York 1ason Charter 1q 76 5—6 Daveio nts in European Public Rita Networks Proc Conf on Computer Networks arid Teleprocessing Aachen 1976 KIRSTFIN PT 38—60 KIRSTEIN PT Planned New Public Data Networks Computer Networks Vol 1 No 2 Sept 1976 HIG NSON PL Software Requi ren n t s of itewav between Networks ed T R Bell and 197t 257—67 S’IOKES AV i nt c i uter System for Use as a Minicomputer Software i n- c Bell Ar terdani North—Holland S ttwar BevolopTent in a Network Environrrent usinci a h hiqh—L vel Lanquacje -tin icc m U t e r coftW Lro ed J R Bell and P 2 Bell Amsterdam North—Holland 1976 241—55 STOKES AV KI ThLN VT mu i tori ng and Acess Control of the Uandon Ncx3e of AR°ANE’I Proc Nat Ccxnputer Conf 4 5 1 9 7 6 597—603 - ——-- - - _ _ _