THINK Home Digital Archive Search Feedback




A Technical History of the ARPANET -
A Technical Tour


Overview of Host-to-IMP protocols

Details of Host-to-IMP protocols

Overview of the Host-to-IMP Protocols

"With IMP-HOST protocol, a HOST has operating rules which permit it to send messages to specified HOSTs on the ARPANET and to be informed of the dispensation of those messages. In particular, the IMP-HOST protocol constrains HOSTs in their transmissions so that they can make good use of available communications capacity without denying such availability to other HOSTs" [Crocker 1972, "Function-oriented..."]

The Network Working Group (NWG) used a layered approach to the specification of the communication protocols. [Completion Report, III-60] The higher layers, such as programs like telnet and FTP would make use of the host-to-host layer, which would in turn make use of the host-IMP layer. The hosts never see the IMP-IMP protocol implementation and have no need to know about the inner workings of the IMP subnetwork. This layered approach permits later development of ad hoc protocols at the top of the layered protocol stack. [Completion Report, III-62] (See figure below)

The way the host and IMP communicated back-and-forth was considered the host-IMP protocol.  The host would pass its messages to its IMP that were to be delivered to another host on the network. The destination IMP would then send acknowledgements back to the host on the status of the messages. If a message was successfully received by the destination host, the destination IMP would send a RFNM (ready for next message) and if a message was lost in the network the destination IMP would send an incomplete transmission message. The destination IMP was responsible for the reassembly of the received packets into the original message and sending that fully assembled message to its host. An IMP was also able to block incoming messages from its host for various reasons.


Explore the details of the Host-to-IMP protocols

  • Protocol Data Unit - Variable length messages with a max size of 8096 bits. This is the same as the Host-to-Host message, with the exception of control messages sent from the Host to its attached IMP.
  • Addressing - A Host only has one IMP that it can communicate with, the local IMP. Therefore, no addressing is required for Host to IMP communication.
  • Multiplexing - The message ID was used for multiplexing several independent data streams.
  • Fragmentation and Reassembly - None at this layer. All fragmentation and reassembly occurred at the Source IMP-to-Destination IMP layer.
  • Routing - Not applicable. There is only one route between a local Host and a local IMP.
  • Communication - Details
  • Error Detection/Recovery - Details
  • Flow Control - Details
  • Congestion Control - Details




Written by the THINK Protocols team, CS Dept, UT Austin
Please direct comments to Chris Edmondson-Yurkanan.