Skip to main content

III. Conclusion

Extensions to the Protocol​

Some issues have not been adequately treated in the current protocol. We have in mind the following topics to consider more thoroughly and perhaps experiment with.

  1. More Sophisticated Flow Control.

    As mentioned above, other schemes for flow control are still being considered. Other than the necessity of providing some form of it, we are completely unsure of the nature of the problem. It may turn out that the present scheme is completely adequate; it may also turn out that we will need a much more complex scheme.

  2. Error Detection and Recovery

    As we gain some experience with the network, we will develop a better understanding of what errors can occur and, perhaps more importantly, what to do about these errors. We expect the protocol to change as we understand error control.

  3. Start Up and Shut Down Procedures

    We have not done enough thinking about the problem of the host which participates part-time in the network, which ceases normal network operation but remains on the network for special purposes, or which recovers from a system failure. These issues are critical to robust network operation and are possibly our highest priority.

  4. Query and Response

    A host-to-host status test would be a valuable tool, but it is not yet clear what is appropriate to provide.

Coming onto the Network​

We suggest that hosts come onto the network gingerly. First, each host should thoroughly exercise connections to itself. Then it should arrange experiments with some other host who is already functioning. Finally, it may begin to exercise the facilities of other hosts. It is not clear at this time which host will be in the best position to help other hosts first, but UCLA will attempt to serve this function.

Private Networking​

A common ploy is to use the IMP to connect several local computers, one or more of which is not available to the whole network. For example, Harvard is connecting its PDP-1 to its PDP-10 via an IMP; Lincoln Laboratories is connecting its TSP to the 360/67 and the TX2 via an IMP; and UCLA is similarly connecting a XDS 920 to its Sigma-7. In each of these cases, the small machine will not initially provide services to the network.

Although there should be no unwanted traffic to any of these extra hosts, it is desirable that they conform minimally to the network protocol. Provided that they never initiate a connection or send out spurious control commands, it is sufficient for a host to respond to CLS commands with acknowledging CLS commands, and to respond to ECO commands with ERP commands.