The Interface Message Processor for the ARPA Computer Network1
F. E. Heart / R. E. Kahn / S. M. Ornstein / W. R. Crowther / D. C.
For many years, small groups of computers have been interconnected in various ways. Only recently, however, has the interaction of computers and communications become an important topic in its own right.2 In 1968, after considerable preliminary investigation and discussion, the Advanced Research Projects Agency of the Department of Defense (ARPA) embarked on the implementation of a new kind of nationwide computer interconnection known as the AIRPA Network. This network will initially interconnect many dissimilar computers at ten ARPA-supported research centers with 50-kilobit common-carrier circuits. The network may be extended to include many other locations and circuits of higher bandwidth.
The primary goal of the ARPA project is to permit persons and programs at one research center to access data and use interactively programs that exist and run in other computers of the network. This goal may represent a major step down the path taken by computer time-sharing in the sense that the computer resources of the various research centers are thus pooled and directly accessible to the entire community of network participants.
Study of the technology and tariffs of available communications facilities showed that use of conventional line switching facilities would be economically and technically inefficient. The traditional method of routing information through the common-carrier switched network establishes a dedicated path for each conversation. With present technology, the time required for this task is on the order of seconds. For voice communication, that overhead time is negligible, but in the case of many short transmissions, such as may occur between computers, that time is excessive. Therefore, ARPA decided to build a new kind of digital communication system employing wideband leased lines and message switching, wherein a path is not established in advance and each message carries an address. In this domain the project portends a possible major change in the character of data communication services in the United States.
In a nationwide computer network, economic considerations also mitigate against a wideband leased line configuration that is topologically fully connected. In a non-fully connected network, messages must normally traverse several network nodes in going from source to destination. The ARPA Network is designed on this principle and, at each node, a copy of the message is stored until it is safely received at the following node. The network is thus a store and forward system and as such must deal with problems of routing, buffering, synchronization, error control, reliability, and other related issues. To insulate the computer centers from these problems, and to insulate the network from the problems of the computer centers, ARPA decided to place identical small processors at each network node, to interconnect these small processors with leased common-carrier circuits to form a subnet, and to connect each research computer center into the net via the local small processor. In this arrangement the research computer centers are called Hosts and the small processors are called Interface Message Processors, or IMPs. (See Fig. 1.) This approach divides the genesis of the ARPA Network into two parts: (1) design and implementation of the IMP subnet, and (2) design
1Proc. AFIPS SJCC, 1970, pp. 551-567.
2A bibliography of relevant references is included at the end of this paper; a more extensive list may be found in Cuadra .
previous | contents | next