MEMORANDUM RM-3638-PR AUGUST 1964 ON DISTRIBUTED COMMUNICATIONS IV PRIORITY PRECEDENCE AND OVERLOAD Paul Baran PREPARED FOR UNITED STATES AIR FORCE PROJECT RAND we 2L an SANTA MONICA CALIFORNIA MEMORANDUM RM- 3638 -PR AUGUST 1964 ON DISTRIBUTED COMMUNICATIONS IV PRIORITY PRECEDENCE AND OVERLOAD Paul Baran This research is sponsored hy the I nited States Air Force under Project tract No AF 49 638 -700 monitored by the Directorate of Development Plans Deputy Chief of Staff Research and Derelopment Hq USAF Views or conclusions contained in this Memorandum should not he interpreted as representing the oHiCial opinion or policy of the United States Air Force DDC AVAILABILITY NOTICE Quali ed requesters may obtain copies of this report from the Defense Documentation Center DDC 74c Deosz 1700 MAIN ST - SANTA MONICA - CALIFORNIA - vows Copyright 1964 THERANDCORPORATION PREFACE This Memorandum is one in a series of eleven RAND Memoranda detailing the Distributed Adaptive Message Block Network a proposed digital data communications system based on a distributed network concept as presented in Vol I in the series Various other items in the series deal with specific features of the concept results of experimental modelings engineering design considerations and background and future implications The series entitled 0n Distributed Communications is a part of The RAND Corporation's continuing program of research under U S Air Force Project RAND and is related to research in the field of command and control and in 'governmental and military planning and policy making The present Memorandum the fourth in the series is concerned with the establishment of traffic precedence doctrines designed to achieve optimum utilization of the communications resource especially within a seriously degraded and overloaded network The proposed all-digital network has properties that differ in many respects from conventional communications networks Network control features that can be performed only with great difficulty in conventional systems can be readily incorporated into a network of the type contem- plated While this Memorandum is oriented primarily A list of all items in the series is found on p 61 around the distributed digital network the considerations are sufficiently general to be of interest to those con- cerned with the best utilization of an overloaded and impaired command and control communications network -V- SUMMARY In standard military communications systems there is a natural upgrading of message precedence in times of stress But during the 99 per cent of the time of little or no stress conditions the circuits are not conditioned for this high-stress traffic The Distributed Adaptive Message Block Network system delivers all traffic within the same time constraints and that amount of time is within the limits of present emer- gency-level requirements The circuits are prevented from overloading only by restriction of input--and no complete cutoff of input occurs to any user This Memorandum considers four separate techniques that can be used singly or in combination to achieve through automation best use of a seriously degraded and overloaded communications plant within the framework of a rapidly changing organizational structure In the schemes considered precedence is determined moment-by-moment automatically for all traffic in the network Precedence is computed as a composite function of l the ability of the network to accept additional traffic 2 the importance of each user and the utility of his traffic 3 the data rate of each input transmission medium or the transducer used 4 the tolerable delay time for delivery of the traffic _vi_ During overload conditions precedence status information is fed back through the network to limit the type and _volume of data flow allocated to each user The dual goal is to prevent complete denial of com munications to any network user while preventing network overloading The surviving data rate however meager under heavy network degradation is always equitably rationed among the many network users The definitions of equitable and important change from time to time with control reserved to the commanding authority The examples used for describing the proposed notions are based upon the use of a high data-rate time division distributed system whose flexibility permits a rapid ex- change of channel allocation between many users using few bits per second and a few users using many bits per second While the automated computational apparatus needed could in all probability be practicably implemented only in a future communications network the concept is quite general and is easily visualized as being applicable to a pile of papers on a busy executive's desk to a stack of computer programs awaiting processing or to any other form of communications system -vii- ACKNOWLEDGMENTS I am indebted to Jack Carne for many discussions that raised some of the notions included in this work would also like to acknowledge the continuing aid of Keith Uncapher and Wade Holland in the preparation of this series The parallel but independent work of Marvin Adelson at System Development Corporation is also to be noted In an internal SDC working paper he has briefly suggested a somewhat related direction to the common problem I would also like to acknowledge a conversation with Arthur Rosenberg of SDC on the subject and to mention the excellent work of Col A J Mandelbaum appearing in reports of the Stanford Research Institute describing military communica- tions traffic overload problems I would also like to acknowledge many always stimu- lating discussions with R H Scherer of the Office of the Director of Defense Research and Engineering on com munications and in particular his statements on the dif ferences between perishability and the importance of military traffic C B Laning of System Development Corporation Marvin Adelson now at the National Academy of Sciences Jack Carne and James Farmer of RAND reviewed this manuscript and made many excellent suggestions which have been incorporated Mandelbaum A J Speed of Communications Precedence and Traffic Control U Technical Report 2 Stanford Re- search Institute Project 2841 Menlo Park California Inasmuch as the thoughts raised in this paper are highly Speculative and may be regarded by some as flights of fancy beyond the realm of justifiable experience I alone must take the blame CONTENTS 0 000 000 0000ACKNOWLEDGMENTS vii Section 00000000000000000 1 II TRAFFIC OVERLOAD 3 Store-and-Forward Overload 5 Line-Switching Overload 7 THE DISTRIBUTED ADAPTIVE MESSAGE BLOCK NETWORK 8 Hot-Potato Routing 9 The Line-Switching Illusion lO Avoiding Network Overloading-- Choking 12 The Monetary Analogy 17 IV USER CONTROL 19 The Communications Control Console 19 An Analog Model of the Console 24 V MEDIA CONTROL 27 VI PERISHABILITY CONTROL 38 VII CONCLUSIONS 47 Appendix A DCS SPEED OF SERVICE CRITERIA 49 B MINIMIZE--SELECTED EXCERPTS FROM AIR FORCE REGULATIONS 50 C COMMERCIAL TELEPHONE TRAFFIC OVERLOAD PROTECTION TECHNIQUES 54 D DIGITAL COMMUNICATIONS MEDIA 58 LIST OF PUBLICATIONS IN THE SERIES 61 I INTRODUCTION Simulation studies of switched communications network vulnerability such as the RAND NATCOM model invariably include the assumption that any surviving link can carry all the traffic generated in the residual network Since such vulnerability models have been so widely adopted it is appropriate to consider some of the implications of the effects of overload on communications systems that are often ignored in the interest of simplification In this Memorandum we briefly touch upon the over- load problem in present day networks reserving the bulk of our comments for improvements to a future system Though this may appear to be ignoring the present in favor of a system that does not exist the suggestions raised are intended primarily for a far-future time-frame The general principles however might be used in such present- day problems as determining the order of program handling in a time-shared command-control computer In the process of describing the methods for automat- ing the handling of mixed traffic of different levels of importance data rate and priority simple mechanisms are mentioned which create an illusion of mechanized judgment Nothing magical is proposed merely that a Eldrige F R The Effectiveness of Command Control in Strategic Operations for the Mid-Sixties U The RAND Corporation RM-3152-PR October 1962 Secret Reinertsen R W The IBM 704 Computer Communications Vulnerability Model FOUO The RAND Corporation S-l35 August 15 1960 small set of executive policies can be automatically exe- cuted with retention of the right to rapidly change the weights of these policies by human intervention In essence management's fundamental law of exception is mechanized so as to unburden humans processing information--nothing more We are dealing with two separate types of overload terminal and network If a called party is busy on the telephone it may be said that the call is not completed because of an overload at the terminal If the call is blocked by an overload at an intermediate switching center or all circuits are busy then the network is said to be overloaded There are two general categories of communications networks store-and forward and line-switched A torn- tape telegraph switching center which stores messages until a desired circuit is Open is an example of the store-and- forward network The conventional telephone network which closes switches to provide a real-time path from user to user is a line-switched network The overload phenomena have much in common whether they be terminal or network in store-and-forward or line- switching networks and can be discussed on a general con- ceptual basis II TRAFFIC OVERLOAD A communications network can carry only a finite volume of traffic if this volume is exceeded a degrada- tion in performance will result If a local step-by step civilian telephone office is overloaded and unable to ac- cept a call the busy signal is immediately heard If the terminating central office is busy the busy signal is returned shortly after the number is dialed Usually this delay is annoying but not critical One can wait and try again Commercial communications networks are designed around an underlying assumption that each telephone is probably used less than two per cent of the time A small amount of switching equipment can be safely shared by a large number of intermittent users It has been found historically that telephone sub- scribers are well satisfied if they are able to pick up the telephone and obtain service within ten seconds 99 per cent of the time This level of service is called the probability time P-T 0 01 10 that is the probability that service will be unavailable after ten seconds will occur only about one per cent of the time for any subscriber An additional doctrine of commercial common-carrier switched networks is that every subscriber shall receive the same grade of service R I Wilkerson's comments to M Juncosa and R Kalaba in Optimal Utilization of TrunkingiFacilities Communications and Electronics No 40 January 1959 p 1002 Ninety-nine per cent service may sound like excellent communications service But to the military user such a network leaves much to be desired The user has no choice as to exactly which one per cent of the timef service is to be denied The one per cent failure time must be expected to occur when there is an abnormally heavy demand upon the network In civilian networks this usually happens during unexpected snowstorms widesPread fires hurricanes floods etc This can be tolerated since the goal of the commercial telephone utility is to provide service at the lowest cost mg _of the time it is not basically intended for general emergencies Military crises almost by definition place abnormal loads on systems When using a communications system de- signed under such civilian loading assumptions for military purposes one can expect most service denials to occur precisely during those times when most needed Though this may sound like a recitation of the obvious the im- plications of such underlying ground rules have on many occasions been unappreciated by some planners of systems for the military This common implicit system-design assumption is particularly treacherous as there is little opportunity to discover its existence under normal test operation munications networks are rarely exercised in real-time to simulate extensive communications network damage and overload STORE-AND-FORWARD OVERLOAD Overloading is one of the causes of breakdowns in store-and-forward systems during military crises During periods of high tension not only does command-control traffic increase but even logistics traffic is generated in greater volumes In a crisis almost everyone feels obliged to communicate the heavy increase of low-priority logistics traffic has been called the underwear ordering effect The crisis will evoke a flood of backlogged requi sitions into the system all demanding immediate processing A significant increase can usually be accepted by the local communications tributary station for processing the bottleneck occurs farther Present-day hard copy writtenwtext military communications networks are slow-speed store-and-forward systems Long-time inter- mediate storage is used at the switching nodes to improve high-cost long-line-circuit usage When the traffic volume arriving at the intermediate switching center from the many feed points is greater than the output circuit can handle messages must be backlogged There are several different military precedence systems in use which theoretically insure that the more urgent traffic is processed first For example in the BIX Binary Exchange system the precedence cate- gories of Flash Emergency Operational Immediate Priority Routine ' and Deferred are used These The ex post facto explanation heard from the requi sitioner is If war is coming up we better get all our requisitions in the mill so we will be ready to fight categories are divided into two subsets High Precedence and Low Precedence each handled by a different set of rules within the store-and-forward switching center The earlier UNICOM system originally considered only three grades of precedence Priority and Routine more recently this program has incorporated the use of four precedence grades The Defense Communica- tion Agency has standardized on the set shown in Appendix A Each user of the conventional military store-and- forward system is reSponsible for suitably marking his own messages fed into the network The military communicator s goal is to process all highest priority traffic so that it is delivered to the addressee within minutes of trans- mission all of the next grade traffic within minutes nL 1 etc Since these times cannot be met under heavy traffic conditions a downgrading procedure called Minimize is called into play see Appendix B Minimize attempts to reduce the volume of high-precedence traffic Although helpful the procedure has not been wholly effective in reducing network overload The writer has heard it said that the following chain of events occurs Messages in- serted into the network are delayed enroute by a time factor unknown to the originating party Not having any confirmation of receipt of his urgent message the orig inator panics and sends another message--at an increased precedence level Still not receiving an answer the originator again panics and dumps still more high-precedence traffic into the network The values of X Y etc apparently are not stan- dardized for different systems This mechanism may be viewed as being akin to a fast- acting servo controller connected in a loop with a long signal feedback lag-time oscillation can be expect d A communication network that does not let the user know how long it will be before his message will be delivered to the end addressee may be theoretically oscillatory LINE-SWITCHING OVERLOAD A similar overload mechanism is sometimes observed in the civilian telephone system during unexpected civil crises A caller dials a number and hears a busy signal Upon receipt of the busy signal he dials again Again hearing the busy signal he becomes more impatient and once more ties up the shared central office equipment In extreme cases the telephone company may evoke a doctrine called line-load control This consists of intermittently cutting off blocks of subscriber lines feeding the over- loaded offices Line-load control not only provides better service to the remaining users but safeguards against a complete central office failure by overload of the time- shared common-control markers used in some dial systems Such panic-induced failures tying up an entire central office for several hours have been noted The telephone utility in its efforts to maximize continuity of service in the face of overload calls a series of separate tech- niques into play as the load builds up near the critical point these are described in Appendix C THE DISTRIBUTED ADAPTIVE MESSAGE BLOCK NETWORK In order that a military communication network die gracefully it must overload gracefully The proposed broadband distributed network described in this series has several orders of magnitude greater communications capability than any existing military network Although overloads would not normally be expected in such a network we must still plan for such an eventuality In the Distributed Adaptive Message Block Network all traffic from each originating circuit is chopped into small blocks of data called Message Blocks Each Message Block is rubber stamped with the symbols of the end destination the originating station and some other house- keeping data These Message Blocks'are transmitted from Switching Node to Switching Node eventually reaching the desired end station by a reasonably efficient path In order to handle real-time digital voice transmis sion it was necessary to limit the amount of in-transit storage at each node to minimize differential-path delay times Because of this restriction on the amount of storage capacity at each node consideration was limited to only those routing doctrines that utilized little storage at the Switching Nodes While such doctrines were expected to be less efficient in line utilization than the doctrines which allow backlog of many bits at each switch ing center under simulation it has been found that little circuit utilization performance is lost of the high line-utilization of store-and-forward can be had with 3251 little in-transit storage By limiting the amount of storage the store-and forward system can be made to ex- hibit one of the most useful features of a line-switching system avoidance of store-and-forward oscillation by immediate confirmation of receipt of message This is accomplished while retaining most of the store-and-forward system's advantage of greater alternate-routing capability and equitable sharing of a single channel by a plurality of users HOT-POTATO ROUTING In the distributed network routing scheme under con- sideration the policy is used that if the preferred path is busy the in-transit Message Block is not stored but rather sent out over a less efficient but non-busy link This rapid passing around of messages without delay even if a secondary route must be chosen is called a hot- potato routing doctrine Each node tries to get rid of its messages as if they were hot potatoes and the node is not wearing gloves With such a doctrine only enough storage at each node need be provided to permit retransmitting messages if an acknowledgment of correct receipt is not received from the adjacent station within a prescribed time interval Message storage capacity is modest For example if eight separate full-duplex 1 5- megabit sec 150-mi links feed a single Switching Node only about 32 000 bits of storage are required at each node Present estimates are that it will take on the order of one millisecond for a 1024-bit Message Block to be read into a node for a decision to be made as to the best direction in which to route the Message Block and to start the Message -10 THE LINE-SWITCHING ILLUSION Although the broadband distributed network is de- scribed as a store-and-forward system which it is as far as internal network operation is concerned the fast switching time of the hot potato doctrine presents the illusion of a virtual circuit having a short delay be- tween two users While it is anticipated that this mechanism will eliminate the open servo loop store-and forward delay problem discussed above this is only part of the problem Figure portrays several Message Blocks arriving at a distributed network Switching Node Each node contains a routing decision mechanism Messages are shown to arrive simultaneously from different lines Three of the mes- sages--those coming in on Lines 1 3 and N --are all to be delivered to Station A Line 2 carries a message directed to Station D while Lines 4 and 5 carry messages which are to terminate at Stations and F reapectively Block on its way to an adjacent node Thus if a Message Block has to traverse 30 such nodes in going across the U S only about 30 milliseconds will be required to pass through 30 Switching Nodes plus another 20 milliseconds transit time through 4000 mi of transmission lines at 133 000 mi sec Thus we anticipate that it may take only about 50 milli- seconds to transmit data across the U S This is less than the delays encountered in forming digital voice into Message Blocks and unpacking the Message Blocks into audio voice This subject is discussed in detail in ODC-VII CDC is an abbreviation of the series title On Distributed Communications the number following refers to the particu- lar volume within the series A list of all items in the series is found on p 61 -11- Shortest path to Station A Fig l--Message Blocks Arriving at a Switching Node -12- A dynamically-updated routing table stored at each node indicates the best route for each Message Block to take to reach its directed end terminal station In the example three separate messages are received simultane- ously at the node each wishing to take the best path toward A The shortest route is via Line 2 but all three messages cannot be sent out simultaneously so two of the messages must travel over alternate paths Although we could have built a system that stored the currently un- deliverable messages until the line to A cleared we have chosen to limit ourselves to a system which sends the overflow messages over less efficient paths effectively preventing a common type of overload see ODC-II When two messages seek the same preferred line a random choice is used to select which message is sent out over the best path In other words if three signals are all to be directed towards A it is not felt that it makes much difference whether the first choice second choice or third-choice path is taken since only a few extra nodes must be traversed before the end station is reached Simulation has shown that this use of secondary paths in lieu of storage is a surprisingly effective doctrine enabling transmission of about 30-50 per cent of the peak theoretical capacity possible in a store-and forward system having infinite storage and infinite delays AVOIDING NETWORK 0VERLOADING--CHOKING Choking is a policy automatically executed by the computer logic at each Switching Node to cut off new local traffic to a network approaching overload A sample of -13- such a choking doctrine is that shown in Figs 2-4 and described below Each node as shown in Fig 2 has input lines feeding traffic into the node and output lines all carrying traffic away from the node In the example 8 1 There is an input line and an output line for each remote station 2 Ii input data rate for the i'th line i 3 0i output data rate for the i'th line i 4 Compute 7 8 Eli- 201 i 1 i I which is the backlog traffic passing through the node should always be a negative number-- otherwise more traffic is piling up than is being carried away 5 Compute where is the volume of locally- generated traffic that may be allowed to enter the network If all or almost all of the output lines are busy the local node avoids feeding new traffic into the network until there is open network output capacity Implementa- tion of this simple intermittent-connection choking doctrine in the illustration presents a varying network input ca- pacity to the tributary communications center feeding Line 1 proportional to what the network can comfortably carry Net- work simulation described in ODC-II and confirms that -14- r0 REMOTE SWITCHING LN OUTPUT INPUT LINES MODE LINES I 1 I I I I MEASURE THROUGHPUT I I LOAD CONTROL LOCAL OUT TRAFFIC Fig 2--Choking Input Traffic to Switching Node to Prevent Network Overloading -15- a simple policy that chokes off input signals in a binary manner is sufficient to protect any station or combination of stations from overloading the network while allowing each station to retain a high input rate Figure 3 represents a variable volume of traffic passing through a sample node Each of eight input lines is assumed to be able to carry a peak capacity with the following sample choking doctrine im- posed If the total traffic in the network approaches a volume greater than a preset value 223 input will be pre- vented from entering the network Traffic already in the network is given preference over newly-entering traffic Once traffic has entered into the network the user is almost guaranteed delivery with the certainty that his own traffic will not cause a local bottleneck at another switch- ing center and get bogged down in transit In Figs 2 and 3 the number of lines in simultaneous use is measured to provide a rapid estimate of those in- termittent periods when new traffic can be safely entered into the network The measuring detector in this case might be a simple majority weighting element reSponding in this case to a threshold of five out of eight If the traffic volume is less than threshold the full input traffic capacity at the link is allowable But when the nodal threshold is exceeded the input is cut off in a binary manner as shown ianig 4 The local user sees these very short on-off periods smoothed out with the network input capacity appearing as a slowly varying allowable data rate the maximum input rate that can be fed into the network without causing an overload NODAL THROUGHPUT 500 000 -16- Total traffic passing through node Short term transient Fig 3--Traffic Load Through Sample TIME Node 0 8 NODAL INPUT l 500 000 SEC Long time average allowable traffic that may be input at local node ON OFF on on ow TIME ml- Fig 4--Intermittently Connected Local Input to Limit Average Local Data Rate -17- The next question we shall consider is allocating this total locally-allowable data rate among a large set of local users or subscribers THE MONETARY ANALOGY Each local communications center in a communications network can only feed up to a certain number of bits per second into the network without overload This total variable volume may be thought of as being a communica- measured in bits rather tions resource or a currency ' than dollars This currency can theoretically be Spent in any manner desired At present in military networks we generally allow each user to demand as much of the limited resource of communications capability as he alone desires limited only by the precedence of his traffic During times of crisis military commanders will try to Spend more communications currency than exists and there will be an effect identical to inflation Messages once labeled Deferred will be stamped Operational Immediate and jammed back into the input hopper It is analogous to the inflationary competition of competing buyers for scarce goods We temporarily delude ourselves into think- ing that we are buying more capability by inflating the precedence indicator It is only when we are forced to face up to reality by a currency devaluation that we ap- preciate what has happened The Minimize doctrine described in Appendix B can be seen to be analogous to a currency devaluation In designing a priority handling system we should never permit ourselves to believe that -18- we have more or less usable communications capability than we really have This implies network status control feedback loops -19- IV USER CONTROL A tributary communications station is a traffic con- centrator Messages or lines from many individual users are combined at the communications office thus making most economic use of an expensive and limited communica- tions facility through time sharing Although communications centers of the future can be smaller and more completely automated we would like to retain human judgment to allow redistribution of the com- munications resource during crises THE COMMUNICATIONS CONTROL CONSOLE The job and the equipment we describe does not exist today At each local communications center we propose there be a responsible military officer who is charged with local traffic control and who is responsible for changing the gross allocation of the communications re- source when necessary The Communications Controller is to be an arbiter with the task of allocating the available data capacity or communications resource among the many local users of the communication network To aid the Communications Controller to allocate surviving capability equitably among individual demands a Communications Control Console is prOposed This con sole provides a picture of both the availability and the demand for communication service together with devices for its sub-allocation The commander views his communica- tions capacity resource observing present allocation -20- of currency and the size of individual demands spending for service Figure 5 is an illustration of a hypothetical Com- munications Control Console A portion of this console is shown in more detail in Fig 6 The meter in Fig 6 indicates the summation of total demand by all local sub scribers of channel capacity in the common measure of bits per second The right-hand meter indicates the network's ability to absorb new traffic in bits per second after choking Since traffic already in the network is given precedence over new traffic overload is prevented by the diminution of acceptance of new traffic Thus the total input data rate from all local users into the network must be limited until an equilibrium point is reached between the traffic that is admitted into the network and the traffic that the network can rapidly deliver Each communications center need not use its full allocation of capability for feeding traffic into the network Conversely a commander at a particular remote site at a crisis point may handle a heavier-than-normal communications load--if necessary by requesting nearby stations to reduce their traffic input in order to allow heavier local use of the network capa- bility Flexibility is thus reserved to allow for the many different centers of various importance which utilize the network Further each individual network user subscriber can ask the Communications Controller for more than his I This is the Priority Control Console described in detail in Green light indicates demand less than indicates allocated data rate Signed copoc'iy for each separate military function or organization Red light indicates demand is approaching assigned capacity '99 oo 0 Indicates summation of total demand by all local subseribers bits second 0 om Indicates netw0rk ability to absorb new traffic bits second 0 9 1 Set to limit total locally generated traffic bits second Master anel nit panels Set to limit traffic to each separate military function or organization bits second See cletailI Fig 6 Fig 5-The Communications Control Console Priority Control Console -21- -22- NEAR OVERLOAD 1 a 3 DEMAND ALLOCATED 50 60 4 80 90 ASSIGN Fig 6--Detail of Communications Control Console Traffic Mbnitoring and Control Unit See Fig 5 -23 currently allocated share of the communications resource This change of assignment is effected by the Communica- -tions Controller turning a dial labeled Allocated in Fig 6 This control sets the limit to the total volume of locally-generated traffic in bits sec If for example a station was requested to cut its loading to 50 per cent of its allocated capacity the dial would be set accordingly In Fig 5 two lights are shown above each panel The green light left side indicates situation normal demand safely under the allocated capacity The red light right side warns that traffic volume is approaching the overload point Each local Communications Controller in turn can ration his assigned total communications resource among his many competing local users For the sake of illustra- tion assume that each user performs a vastly different function for example intelligence early warning logistics supply etc To allow control of these separate further- subdivided functional categories individual panel units similar to the master panel are provided The first such unit might control the gross volume of intelligence traffic the second early warning traffic etc Thus the Communica- tions Controller would also have at his fingertips the ability to vary the allocation of data rate to each separate function The Communications Controller will not play the con- sole as an organ since gross changes in loading are slowly- occurring phenomena rather he will normally leave the controls set to fixed positions except when a crisis or overload approaches as indicated by the red warning light -24- He then decides which users with growing demands should deprive others with less important duties and to what degree Every user who desires more communications capa- bility than he is momentarily allocated is always allowed to talk to the Communications Controller to present his case for an increase in allotment of bits The console does not seek to supplant human judgment It simply provides an automated facility to instantaneously implement the human executive decision It administers the will of the commander swiftly and without an arbitrary cutoff No individual function need ever be totally de- prived of instantaneous communication as is necessary with the binary decision rules necessary to enforce today's precedence doctrines AN ANALOG MODEL OF THE CONSOLE Figure 7 is a sketch of a schematic analog-signal model of the Communications Control Console This hypo- thetical implementation is included to suggest that the circuitry required to diSplay the control traffic loading need not necessarily be complex In Fig 7 an AC input voltage pr0portional to local available data rate is in- serted at the terminals in the upper-left-hand corner of the illustration A rotary autotransformer T-l is con nected to the control knob of the unit A second voltmeter V-2 is connected to the output of T-l to Show the reduced amount of data rate that may be utilized locally Each The all-digital version of this console is described in where all signals necessary for the console are included AC input voltage proportional to locall available data rate Individual channels i T2 9 g A ill Voltage proportional to allowable data rate Egg indicates local bit rate demand Signal to limit number of data transducers connected to system Constant current summation of signals whose current is proportional to hit rate for each data source locally connected are Fig 7-Traffic Monitoring and Control Unit Circuitry Sub assignment local channels -25- -26- unit panel is connected to a separate autotransformer Each autotransformer in turn can be connected to lower- level sub-assignment autotransformers These autotrans- formers create voltages proportional to the assigned data rate of each of the individual unit panels Voltmeter indicates the local bit-rate demand This signal can be approximated by a simple constant-current summation of binary control signals whose currents are proportional to the bit rate demand for each data source energized For example if four teletype machines are connected to the system four separate resistors each passing a current proportional to the data rate required by each are used Each time a teletype machine is energized an associated relay closes attaching an AC voltage to the resistor associated with the connected teletype If all four teletypes are in use the currents passing through all four resistors would be summed producing a voltage whose value would be approximately equal to the total bit- rate demand of all four individual teletypewriter devices 0n the extreme right-hand side of Fig 7 it can be seen that each single-unit channel can be subdivided into other smaller sub assignment panels These panels can be connected to match the same hierarchical organizational structure of the network feeding the local communications transducers to the communications tributary station Nothing of course restricts any user from being con- nected to more than one communications tributary station In practice this would be done at the Multiplexing Station which already has equipment to measure activity -27- V MEDIA CONTROL To this point we have described how available data rate a volatile commodity whose measure is in bits per second may be allocated among the many users of a com- munication network We shall next consider a method to encourage each individual user to make most efficient use of his allocated communications resource The common-user digital network of the future en- counters a wide variety of data-generating devices These devices can load the network at rates differing as much as 100 000 to 1 Some of these devices are briefly described in Appendix D and include telephone teletype telegraph and facsimile Each of these media places a different bit-rate load upon the communications system In the civilian world the decision of choice of medium tends to be primarily economic We ordinarily use the cheapest medium commensurate with the value and perish- ability of the information In the non-dollar economy of the military world we need a mechanism analogous to the concept of money to encourage each individual user to select the medium which makes most efficient use of the total communications resource The reader is cautioned that the following develOpment is not the result of rigorous analysis inasmuch as the analytical tools are simply not available But in their absence we have chosen the following approach the numbers used are not important since their purpose is illustrative only Let us start with a tentative assumption that each alternative communications medium conveys roughly about the -28- same amount of information theory type information per unit time under real-time human-to-human or human to machine information interchange This is not too difficult to visualize For example humans Speak and type at the same rate about 60 words per minute But in the informa- tion theory sense such language is highly redundant Tests have been run to determine the non-redundant information input rate of humans and an extensive body of literature has develoPed examining the input data rate for man's input senses The results of these tests indicate that the maximum possible human input data rate is probably less than about 50 bits sec based upon tachistoscopic picture element recognition visual input language text also visual in- put voice auditory input and vibratory Morse code tactile input As information can be sent in a variety of ways we wish to encourage each user always to uSe the most efficient medium During peacetime it might be most efficient to send a letter by facsimile if transmission bandwidth is cheap and keyboard operator's time expensive However during overload the available supply of the com- munications commodity decreases and assuming a free economy the price should change Under these conditions one would like to encourage the use of more-efficient narrower band transmission devices such as say teletype We would like to have a cost table of these different data sources in order to provide a basis for determining the inflation-free price of service when the communications resource is to be rapidly rationed see Table -29- Table I DATA RATES FOR DIFFERENT DIGITAL GENERATING DEVICES No Bits Sec Bauds TELETYPE 5-Unit 60 Fieldata 80 VOICE 19 2 kilobits 19 200 2 4 kilobits 2 400 FACSIMILE 9 6 kilobits 9 600 DATA Keyboard 40 300 c m Card Reader 5 000 High-Speed Card Reader 20 000 Partial Core Dump_L10Z 100 000 Entire Core Dump 1 000 000 Magnetic Tape Transfer 200 000 -30- Data rate alone however does not provide a complete measure of network loading some devices have a short duty cycle such as one computer sending the contents of its core to a remote computer While such devices place a heavy peak demand for service they are highly intermittent 0n the other hand a pulse-coded telephone call places a lower peak demand load but ties up network capacity for a longer period and results in heavier average loading Therefore we should include an expected message-duration or holding- time factor in the network-load weighting table Two separate factors are at work here Many separate low-data-rate devices time-shared or concentrated into a single high-data-rate link permit better averaging as compared to a few users But as many of the high-data-rate users get in and get out fast they have a short holding time This helps the averaging process To be precise in this computation a better understanding of the number of users their use statistics and the network characteristics appears manda- tory and shall be deferred until such information is available It is sufficient for purposes of explaining the concept to use the following tentative rationale in preparing the sample loading chart of Table II Start with the data rate in bits per second of the transducers which we have briefly considered Next give credit to those devices expected to be in use for only a comparatively small portion of the time Column 1 lists the peak data rate in bits sec for each of the input devices Column 2 lists the percentage of network users using any single type of data input device If for Table II APPROXIMATION OF NETWORK LOADINGS FOR DIFFERENT DIGITAL DATA GENERATING DEVICES 1 2 3 4 5 6 Data Rate Bits Sec Bands Per- centage of Users Time in Use Data Rate Users 1 Long Time Average Loading Rank Order Loadin TELETYPE 5-Unit 60 3 100 180 60 2 Fieldata 80 5 100 400 80 3 VOICE 19 2 kilobits 19 200 60 100 1 152 000 19 200 10 2 4 kilobits 2 400 10 100 2g 000 2 400 7 FACSIMILE 9 6 kilobits 9 600 4 100 38 400 9 600 DATA Keyboard 40 10 20 80 8 300 c m Card Reader 5 000 10 1 500 500 High-Speed Card Reader 20 000 20 8 000 4 000 Partial Core Dump 100 000 0 1 100 100 Entire Core Dump 1 000 000 0 1 1 000 Magnetic Tape Transfer 200 000 30 60 000 11 -31- -32- example ten per cent of the links served 40-bit sec key- boards the value of ten per cent would be shown in Col 2 Column 3 shows the expected maximum duty cycle of each input device averaged over a long period A telephone could conceivably be used 100 per cent of the time while it is doubtful that a computer core dump would occur more often than 0 1 per cent of the time Column 4 is the product of data rate and users and provides an indication of the loading demand by each type of input source Column 5 lists the rank order of entries of Column 4 to indicate the type devices that make heaviest average demands upon the data resource Thus for example in our network we will expect many digital telephone users but few computer core dumps This determination of allocation of loading demand can be further refined by including the following two factors First we could use the investment cost of each input device as a factor indicating its importance Thus for example another multiplicative factor could be included to allow for the higher investment cost of computers relative to telephones Secondly we could also use the reciprocal of the total number of each data input device as the metric Using this rationale in a distributed command-control system the true vulnerability is probably some function of the reciprocal of the number of times a specialized input device is repli- cated in the network For example if three special purpose computers are tied together with a communications network destruction of a single computer installation or its communications is correspondingly more serious - than if the complex contains fifty computers sharing the work In the first case there would be loss of oneuthird of the complex's capacity in the second only 1 50 capa- city Thus we might wish to tend to favor the more critical elements as compared to the less unique The metric being deve10ped as a measure of the com- munications resource is a vector having many components To this point we have described a few of the components We shall now consider yet another component--the conven- tional military precedence indicator and how it might be blended into the vector The present-day precedence indi- cator system concept is based primarily upon the Speed of delivery of a message Historically it has probably grown out of the old commercial telegraph tariffs telegram deferred day letter and night letter Column 1 of Table lists present Defense Communica- tions System Precedence Categories together with target processing times Column 3 lists the approximate ratio of these time categories An underlying consideration in the following development is the mixed requirement that while we wish to give priority treatment to the higher- precedence traffic of equal network loading we must also satisfy the goal that we preserve a minimum transmission capability for the lower-precedence traffic Thus in stead of a blanket rule that all traffic of a given prece- dence grade will be transmitted before handling the next lower precedence grade we choose to use the time ratios of these precedence categories to act as a preference weighting factor Table USE OF SPEED OF SERVICE CRITERIA TO ESTABLISH A NUMERICAL WEIGHTING PRECEDENCE FACTOR Precedence DCS Speed of a Precedence Categories Service Criteria Time Ratio Factor Flash 0 1b 1 Emergency 30 min 10 10 Ops Immed 30 min 10-20 20 1 hour Priority 1 hour 20-100 100 5 hours Routine 3 hours 60-160 160 8 hours Deferred 8 hours 160-300 300 Start of next day's business 8See Appendix A bBased upon message length of 125-150 groups 60 3 3 min or 1 20 hr -35- Table IV combines the network loading factors for different digital services shown in Table I together with the numerical values of preference weighting factor derived from Table II Table is included to illustrate certain combina- tions of lower-precedence traffic which can automatically force preferential treatment over those forms of ''high- precedence traffic making extremely inefficient use of network data rate The entries of Table are listed in order of preference This illustrates how we can encourage each network user to make an efficient choice of the form of data transmission regardless of his chosen precedence grade without depriving any low-precedence user from transmitting a small volume of traffic which adds negli- gible loading to the network LONG-TIME NETWORK LOADING AS A FUNCTION OF DIGITAL SERVICES AND NUMERICAL PREFERENCE WEIGHTING FACTOR Table IV Precedence Grade x1 Flash x10 Emergency x20 Opn Immed x100 Priority x160 Routine x300 Deferred 10 11 Keyboard 5-Unit 704 Fieldata Partial Core Dump 300 c m Card Reader Entire Core Dump 2 4-Kb Voice 1200 c m Card Reader FAX 19 2-Kb Voice Mag Tape Transfer 60 80 100 500 1K 2 4K 4K 9 6K 19 10K 24K 40K 96K 192K 600K 160 1 2K 1 6K 2K 10K 20K 48K 80K 192K 384K 1 2M 800 6K 8K 10K 50K 100K 240K 400K 960K 1 92M 6M 1 28K 9 6K 12 8K 16K 80K 160K 384K 640K 1 536M 2 892M5 9 6M 2 4K 18K 24K 30K 150K 300K 720K 1 2M 2 88M 5 76M 18M 1 000 1 000 000 Table RANK 0F ORDER OF SERVICE PREFERENCE Preference Grade x1 x10 x20 x100 x160 x300 Flash Emergency Ops Immed Priority Routine Deferred Hv l Keyboard 5-Unit 704 Fieldata 300 c m Card Reader 7 20 26 38 41 44 Entire Core Dump 11 25 32 43 45 49 2 4-Kb Voice 17 33 37 48 51 55 1200 c m Card Reader 19 2-Kb Voice 31 46 50 60 62 63 Mag Tape Transfer 39 53 57 64 65 66 Partial Core Dump -37- -38- VI PERISHABILITY CONTROL A communications network can deliver more traffic than a recipient can answer We wish to prevent oscillation by feeding back processing status as a function of perish- ability First we shall consider the number of parameters needed to Specify the perishability of messages of various utilities transmitted in a communications net- work Perishability and importance are not synonymous Inasmuch as we pay for data transmission we assume that we must receive some utility when the message is correctly received by the recipient The earlier the message arrives the more useful it will be as represented in Fig 8 Such a message might be a telegraph request for a hotel reservation transmitted while one is rushing out the door to catch an airplane If the message arrives after the time requested for the reservation it has a utility of zero If the message arrives early enough to guarantee a room then it has a relatively high utility While we lack a good metric for communications utility it is sufficient in this discussion to let it be equal to the quantity l-Ppb where is the probability of sleeping on a park bench Perhaps the economist's term disutility would be more fitting- the price for which one would be willing to sleep on the park bench Only two parameters of specification were needed to provide a reasonable measure of the value of the communication as a function of time -39- a Peak value of utilitx at 0 Point A in Fig b Last 21mg the message had any value Point B A second message example illustrated in Fig 9 is am bringing a guest home for dinner tonight here three parameters of Specification would be useful Message RESERVE HOTEL INSTANTANEOUS UTILITY l-probability of sleeping an a park bench IF TIME tl t2 Time Message Date of is Sent Reservation Fig 8- An Example of Utility of Message as a Function of Elapsed Time -40- 69 Message I'm bringing a guest home for dinner tonight INSTANTANEOUS UTILITY OF MESSAGE Dinner Time - TIME Length of time i1 takes 10 go to grocery store and reiurn Fig 9--An Example of Utility of Message As a Function of Delivery Time -41- a Peak value of utility at 0 Point b Last Eimg the message had any value Point c Last Eimg the message had a high value of utility Point C Parameter might correspond to the last time one's wife could drive over to the grocery store and pick up more beer for thirsty guests Figure 10 shows the value of instantaneous utility for three different messages a message to initiate a large long-term project a message to reserve a hotel room and a message saying Merry Christmas I Only three points are needed to approximate the shape of each of these curves We may for example ask the three questions a How important is the message b Ehgg would we $153 the message delivered c Ehgn is the latest time the message would be of any value While it is possible to think of examples that require more than three specification parameters to delineate the instantaneous utility curve three will suffice for our purpose If the communications network knows these few parameters for all traffic in the network it will be able to perform a rather sophisticated control function Con- sider the narrow-funnel neck problem where many communica- tions links terminate at a single individual A com- munications system may have a capability of delivering more messages which need to be processed than a single end individual can handle In our future communications system we seek to automatically inform the sender not only that UTI LETY IMPORTANCE -42r A MESSAGE TO Initiate a large long-term project Reserve a hotelroonn MON 0 Christmas card TIME Fig 10--Instantaneous Utility Curves for Three Different Messages -43- the end addressee is busy but that he has a processing backlog of an expected value of minutes If this predicted time is acceptable to the sender he will do nothing Otherwise he could increase his 'precedence indicator and try again In Fig 11 Messages 101-105 arrive sequentially at an end receiver who requires more time to process mes- sages than the time interval between receipt he is overloaded The question then becomes one of determining the best order for processing these messages For example we could use the conventional first-in first-out pro- cessing method Let us consider the alternative of com- puting the integral of the peak utility of all messages awaiting attention and ordering the messages in descending sequence of the integrals of peak utility Thus the end addressee receiver sequentially picks out what we hope will be the most urgent messages of the set requiring attention If equipment at the receiver observes the length of time it takes to process each message it then can heuristically estimate the time when it will be able to accommodate each stacked message Thus the sender can be immediately informed of the expected time of action Even more important however is the receipt of an im- mediate response that the end addressee will not be able to take action within the zone of utility For example Fig 12 shows many messages with different values of utility all directed to a single human bottleneck Each message arrives with a zero urgency level but the level rises in proportion to the urgency until the item is either processed or the point is reached at which the URGENCY of PROCESSING INTEGRAL of INSTANTANEOUS UTILITY -44- Message IOI Message IOZ Message I03 Message I04 Message i05 TIME Length of time to process a message fl 1 Fig ll--Several Almost-Simultaneous Messages Requiring Servicing by a Single Processor NI J0 SESVSSEIW #0 9N 10 AONHOBH Message I Message 2 Message 3 Message TIME Fig 12-Many Messages with Different Utilities Passing Through a Single Processor -45- -46- message is too old to be of any value Thus the sender can be immediately informed of whether and when he can communicate to the addressee allowing him an opportunity to take alternative action if necessary While only the end addressee ever really knows what is really important this feature would allow a major pre-filtering of mes- sages to be processed by the receiver so that the receiver will be able to spend most time on the mpg important of the important tasks demanding his attention -47- VII CONCLUSIONS A precedence system which is time-dynamic in opera- tion has been proposed The assignment of capacity and capability is designed to be rapidly changed by the military communications controller The goal sought is to automate those procedures which help a communications network overload gracefully under attack The priority order is generally consistent with present military prece- dence concepts except that means are included to encourage the use of lower data rate generating devices in preference to the high-data-rate generating devices in times of overload While we have discussed only hard copy messages the same mechanisms appear suitable for telephone calls The feedback information that the called party will not be able to answer within a prescribed period can be announced by an easily recognizable tone pattern Such signaling tones at the beginning at the mid point and near the end of such rationed calls should also serve to shorten their length Similar tones could also indicate that the calling party is working against an incoming backlog and could indicate the urgency of that backlog All the information needed to implement such a seemingly complex precedence system is available at the Multiplexing Station see in the Distributed Adaptive Message Block Network While the techniques discussed best lend themselves to an all-digital distributed network the general notions apply in any situation where volume-limited communications -43- is found We feel that a usable priority structure might one day be evolved that can be implemented on a semi- automatic basis without extreme continuing demands on executive judgment I The communications control system suggested seeks to implement the law of management by exception Each person has a certain job that he must do as long as he is able to operate within organizational constraints it is not necessary to request help from a higher hierarchical level If the demand exceeds capability only exceptional cases need be transmitted to a higher level for action In the proposed control system changes in network loading are Spotted and diSplayed to predict that a re- construction of allocation of the communication resource may soon be required If the commander concurs he can instantly re allocate his available resource among the many demanders for the resource without complete depriva tion to any We seek to reduce the inflationary struggle for priority treatment of messages by the generators of com- munications far down the hierarchical chain These sources are the sensors which perceive some of the most important information in need of transmission in a crisis At pres ent these individuals do not always have the authority to obtain immediate precedence treatment for their mes sages without a serial series of approvals before a person is reached who has enough authority to place urgent traffic in the network We think we can do better 1961 p 2 Speed of Service Criteria DCA Circular 70-4 312 11 December Maloney E S 001 USMC Defense Communications System Precedence Precedence Prosign Communication Handling Time Remarks FLASH As fast as possible Interrupts lower- precedence EMERGENCY 30 min Ahead of lower- precedence interrupts when necessary OPERATIONAL 0 IMMEDIATE 30 min to 1 hr Ahead of lower- precedence interrupts when appropriate PRIORITY 1-5 hr Does not interrupt lower- precedence in progress ROUTINE 3-8 hr In order received after higher-precedence DEFERRED 8 hr to start of business following day In order received after higher-precedence NOTES 1 Criteria are based on average message length of 125-150 groups 2 Communication handling times indicate elapsed time from receipt at originating communication center to receipt at addressee center DCS SPEED OF SERVICE CRITERIA Appendix A -49- -50- Appendix MINIMIZE--SELECTED EXCERPTS FROM AIR FORCE REGULATIONS 2 What the Program Does The imposition of MINIMIZE conditions is a warning to the users of the facilities and services that it is vital to drastically reduce normal message and long dis- tance telephone traffic so communications directly con- nected with the emergency or exercise will not be delayed 4 Determining Need for Imposing MINIMIZE In the event of an emergency or anticipated emergency the Air Force Communications Message Traffic Control Unit CMTCU will furnish telephone reports to the Director of Communications-Electronics Headquarters USAF on circuit and traffic conditions throughout the Air Force communica- tions system AFR 100-49 5 Authority to Impose MINIMIZE Except as noted below only the Chief of Staff USAF or in his absence his deputy has the authority to impose MINIMIZE either Air Force-wide or for a Specific area or areas However the commander of an overseas major air command may institute MINIMIZE if he has ob- tained prior coordination of the theater commander con- cerned and approval of the Chief of Staff USAF Air Force Regulation lOO-ll Department of the Air Force Washington 15 January 1957 pp 1-4 -51- 6 Procedure for Imposing MINIMIZE When MINIMIZE is implemented Air Force-wide it will be imposed by the words AIR When it is applied to a particular area it will be indicated by designating the area concerned for example Implementation of MINIMIZE by Headquarters USAF will be accomplished by both ALMAJCOM and AFCOMMSTA general messages to insure that all major air commands and Air Force communications centers receive this notifica- tion by the most expeditious means When a commander of an overseas major air command institutes MINIMIZE under the provisions of paragraph 5 it will be accomplished by means of a specifically addressed message to all concerned 9 Administrative Control of Messages Strict control must be established to prevent un- necessary use of electrical communications once MINIMIZE is effected This will be accomplished as follows a Before drafting a message addressed to the affected area or areas the message drafter will satisfy himself that the immediate situation will be adversely affected if a message is not sent b Each message-releasing authority will perform an adequate check to insure that all message traffic released for electrical transmission falls within the categories of traffic authorized in paragraph 10 Except as noted below messages not acceptable for release by electrical transmission will be 10 -52- referred to the drafter for cancellation or consent to dispatch by other means Message-releasing authorities may release for electrical transmission any message traffic which has been delayed to the extent that early receipt of the content by the addressee s is considered vital Messages Authorized for Electrical Transmission Message-releasing authorities will approve for electrical transmission only those messages meeting one or more of the criteria listed below Messages which do not meet these criteria may be dispatched by courier air mail or ordinary mail in accordance with pertinent Air Force directives a b C FLASH or EMERGENCY precedence messages AFM 11-4 REDLINE messages AFR 100 3 INDICATIONS messages AFR 100-4 OPERATIONAL IMMEDIATE precedence messages per taining to USAF wartime capability plans 2 Aircraft movements 3 MANOPED weather messages 4 Casualty messages Messages pertaining to The existing emergency or exercise 2 Joint War Room Annex JWRA operations 3 Emergency Air Staff Actions EASA ll -53- Handling of Messages by Communications Centers and Relay Stations Under MINIMIZE messages will be handled as follows a Disposition of Original Messages on Hand Upon receipt of MINIMIZE Air Force tributary stations will transmit all messages which bear a FLASH EMERGENCY or OPERATIONAL IMMEDIATE precedence All other messages will be returned to message releasing authorities for appropriate action under paragraph 10 Disposition of Relay Tapes on Hand Upon re- ceipt of MINIMIZE Air Force tape relay stations will transmit all messages which bear a FLASH EMERGENCY or OPERATIONAL IMMEDIATE precedence All other messages will be disposed of as pre scribed by USAF Supplement 1 to ACP 127 B Appendix COMMERCIAL TELEPHONE TRAFFIC OVERLOAD PROTECTION TECHNIQUES The fbllowing techniques are either in use by or are being considered by the Bell Telephone Company to minimize system degradation under overload 1 Reduced Inter-sender Timing The length of time spent by a sender waiting for dial pulses from the subscriber's telephone is reduced under heavy traffic conditions For example upon lifting the telephone from its cradle a dial tone is heard If you do not start to dial immediately you needlessly tie up shared common-control equipment Under overload the period that the sender unit waits before assuming that the subscriber is balking is reduced This forces the subscriber to dial his number more quickly 2 Recorded Message When there is an overload in the Direct Distance Dialing system calls can be passed to an automatic play- back device A canned recording announces an overload condition and requests that the caller wait and place his call later A variation on this recording is used in emergency Such recorded messages have proved to be highly effective in the past When people are Specifically requested not to use the telephone unless it is an emer- gency they will generally refrain from doing so -55- 3 Operator Spacing During dial overload a manual operator can cut in to ask for the number being called The Operator will tell the calling party that she will call back later when the line is free This effects a delayed spreading of calls during peak periods 4 Variable Divisions of Circuits Whenever only two circuits are left in a group of trunks between two cities in the Direct Distance Dialing switching hierarchy these two lines are reserved for calls approaching from the higher level of the switching hierarchy This procedure provides precedence for calls that have proceeded a greater distance into the system than incoming calls 5 Network Management On certain peak days such as Christmas network man- agement policies are enforced which prohibit alternate routing A traffic supervisory console is used to prevent round-about alternate routes during periods of heavy network use and to restrict traffic to short efficient routes The decision to invoke such network management is limited to those in the Regional Offices of the system All Regional Control Centers intercommunicate on a single hot-wire loop circuit which permits each center to inform the others of its actions -56- 6 Switching Operator Position During conditions of high overload in which the pattern of demand changes in a predetermined manner on Mbthers' Day for example there is a shortage of opera tor positions in the suburbs and a surplus in the cities It is possible to remotely switch operator positions and transfer trunks by key switches to process local suburban traffic An operator position in downtown New YOrk can be wired to handle calls originating in a particular suburb alleviating overload peaks 7 Line Load Control Line load control as presently practiced divides all subscribers into three separate groups First there are the subscribers who are deemed to merit communica tions in emergencies Such users include police fire mayors neWSpapers editors pay telephone booths etc These users are in turn divided into two groups Each group comprises about 40-45 per cent of all subscribers During emergencies a switch at the Central Office can be closed to deny access to new calls to one group of 45 per cent of the users After ten minutes service is then denied the second 45-per-cent group No one's connection is interrupted during line load control it is simply that no new calls will be accepted from the 45 per cent to whom access is denied 8 Priority It is said to be possible to call the operator in an emergency to identify yourself tell her it is an -57- emergency and expect that she will handle the call in the appropriate manner this is sometimes called Cate- gory Priority within the Bell System -53- Appendix DIGITAL COMMUNICATIONS MEDIA TELETYPE Although a wide variety of teletype codes is in use we will probably most often encounter conventional five- unit and eight-unit start-stop teletype signals These signals are convertible to binary-stream transmission with buffering storage or by sampling at a rate much higher than the length of a single teletype bit VOICE Digital voice transmission can be performed at a wide variety of data rates Seven-sample pulse-code modulation PCM using 8000 samples per second gives excellent voice quality but requires 56 000 bits sec Differential PCM can be performed with 38 400 bits sec Recently this writer heard tests of High Information Delta Modulation HIDM using a data rate of only 19 200 bits sec The quality was good and the intelligibility excellent HIDM uses relatively simple analog-digital-analog conversion equipment and appears to meet the requirements for mili- tary voice transmission including good dynamic range The use of this type of modulation has been assumed in the discussion in the text of this Memorandum Winkler M R High Information Delta Mbdulation International Convention Record Paper 47 3 New York New Ybrk March 28 1963 -59- Another class of digital voice equipment is vocoder equipment In the vocoder voice energy is frequency- division separated by a bank of band-pass filters the output signal strength of each filter is measured and transmitted as a digital signal Vocoder equipment is generally expensive and the reconstructed voice is of low quality Its chief virtue is that it makes extremely efficient use of bandwidth requiring less than 2400 bits sec There are indications that very-high-quality high-intelligibility vocoders might be built in the future having a data rate of 5-10 kilobits sec One development that occurred after the preparation of this Memorandum will change the loading figures for voice by a large factor and should at least be mentioned here In the description of the Multiplexing Station it is pointed out that it is easy to suppress blank Spots in a voice stream without losing quality or breaking This would probably reduce the average voice-conversation data rate when using HIDM for example from 19 200 bits sec to about 5000 bits sec without any drOp in quality FACSIMILE Facsimile transmission normally an analog signal can be converted by conventional analog-to-digital means to operate at about 9 6 kilobits sec COMPUTER DATA A wide variety of manually and semi manually operated data generating devices are used with computers These -60- include typewriter keyboards coded insertion card interrogators Hollerith card readers punched paper tape readers and high-speed teletypewriters These devices are generally characterized by low bit-rate re- quirements In the future we may also find cases where computers might talk to one another Here the amount of data exchanged can be small if only processed data is ex- changed or if humans are involved But there are also applications where it is desirable to exchange raw data between-machines Here the assumption of equal informa- tion per unit time is not applicable because we are not limited by the human factor in such a loop One applica- tion might be a dump or transfer of the major parts of the high-speed core memory of a computer into a remote computer at a high bit-rate--on the order of one million bits sec Magnetic tape can also be read at one station and rewritten into a remote tape unit for later proces- sing perhaps at rates up to 250 000 bits sec -61- ON DISTRIBUTED COMMUNICATIONS List of Publications in the Series I Introduction to Distributed Communications Networks Paul Baran RM-3420-PR Introduces the system concept and outlines the requirements for and design considerations of the distributed digital data communications net work Considers especially the use of redundancy as a means of withstanding heavy enemy attacks A general understanding of the proposal may be obtained by reading this volume and Vol XI II Digital Simulation of Hot-Potato Routing in a Broadband Distributed Communications Network Sharla P Boehm and Paul Baran RM-3103-PR Describes a computer simulation of the message routing scheme proposed The basic routing doctrine permitted a network to suffer a large number of breaks then reconstitute itself by rapidly relearning to make best use of the surviving links Determination of in a Distributed Network J W Smith RM-3578-PR Continues model simulation reported in Vol II The program was rewritten in a more powerful computer language allowing examination of larger networks Mbdification of the routing doctrine by intermittently reducing the input data rate of local traffic reduced to a low level the number of message blocks taking excessively long paths The level was so low that a deterministic equation was required in lieu of Monte Carlo to examine the now rare event of a long message block path The results of both the simulation and the equation agreed in the area of overlapping validity -62- IV Priority Precedence and Overload Paul Baran RM-3638-PR The creation of dynamic or flexible priority and precedence structures within a communication system handling a mixture of traffic with dif- ferent data rate urgency and importance levels is discussed The goal chosen is optimum utiliza- tion of the communications resource within a seriously degraded and overloaded network V History Alternative Approaches and Comparisons Paul Baran RM-3097-PR A background paper acknowledging the efforts of pe0ple in many fields working toward the develop ment of large communications systems where system reliability and survivability are mandatory A consideration of terminology is designed to ac- quaint the reader with the diverse sometimes conflicting definitions used The evolution of the distributed network is traced and a number of earlier hardware pr0posals are outlined VI Mini-Cost Microwave Paul Baran The technical feasibility of constructing an extremely low-cost all-digital X- or Ku-band microwave relay system operating at a multi- megabit per second data rate is examined The use of newly deve10ped varactor multipliers permits the design of a miniature all-solid- state microwave repeater powered by a thermo- electric converter burning L-P fuel VII Tentative Engineering Specifications and Preliminary Design for a High-Data-Rate Distributed Network Switching Node Paul Baran RM-3763-PR High-speed or hot potato store-and-forward message block relaying forms the heart of the prOposed information transmission system The Switching Nodes are the units in which the com- plex processing takes place The node is de- scribed in sufficient engineering detail to estimate the components required Timing calcu- lations together with a projected_implementation -53- scheme provide a strong foundation for the belief that the construction and use of the node is practical The Mnltiplexing Station Paul Baran RM93764-PR A description of the Multiplexing Stations which connect_subscribers to the Switching Nodes The presentation is in engineering detail demonstrat- ing how the network will simultaneously process traffic from up to 1024 separate users sending a mixture of start-stop teletypewriter digital voice and other signals at various rates IX Security Secrecy and Tamper-Free Considerations Paul Baran Considers the security aspects of a system of the type proposed in which secrecy is of paramount importance Describes the safeguards to be built into the network and evaluates the premise that the existence of Spies within the supposedly secure system must be anticipated Security provisions are based on the belief that protec- tion is best obtained by raising the price of espied information to a level which becomes ex- cessive The treatment of the subject is itself unclassified r X QCost Estimate Paul Baran A detailed cost estimate for the entire proposed system based on an arbitrary network configura- tion of 400 Switching Nodes servicing 100 000 simultaneous users via 200 Multiplexing Stations Assuming a usable life of ten years all costs including operating costs are estimated at about_ $60 000 000 per year XI Summary Overview Paul Baran RM-3767-PR Summarizes the system proposal highlighting the more important features Considers the particular advantages of the distributed network and comments on disadvantages An outline is given of the manner in which future research aimed at an actual imple- mentation of the network might be conducted To- gether with the introductory volume it provides a general description of the entire system concept