[ih] Failures of the early Internet

[Noel Chiappa]

    > From: Vint Cerf

    > so I am not sure that single packet floods would have caused a setup
    > delay/congestion unless the ping messages were longer than a single
    > 1008 byte packet?

I had the same reaction you did.

Minor error in your message (which I don't wish to leave in the archive,
un-corrected, lest it confuse some later historian reading this): I think you
meant 'bits', not "byte[s]". ARPANET 'packets' were "approximately 1000 bits"
(1970 Spring AFIPS paper), "1103 bits, including leader" (1822, January 1976
revision). 'Messages' were at most 8160 bits, 1020 bytes (1822, January 1976
revision).

(For explain some ARPANET-specific terminology to those who aren't ARPANET
experts, 'packets' were IMP->IMP packets [in the modern sense of that term],
'messages' were IMP->Host packets.)

	   Noel