[ih] UDP Length Field?

[Stephen Casner]

On Wed, 2 Dec 2020, Vint Cerf via Internet-history wrote:
>
> I have a fairly clear recollection of being at USC/ISI with Jon, Danny
> Cohen and David Reed in 1977 discussing the real-time application needs
> that TCP would not satisfy owing to retransmission delays.

I can recall that meeting at ISI as well, but the details are all
fuzzy.  One memory that stuck with me but that I can't reconstruct
accurately was of Danny at the blackboard (I think it was still chalk
then) talking about the need for a communications service that avoided
the delay imposed by TCP's reliability mechanisms.  He described the
two communication services as functions or transforms written in the
notation of equation, where TCP's service was a composition including
the reliability function R.  He said something like perhaps we could
get the service we wanted by composing the TCP transform with an
unreliability function R (superscript) -1, but concluded that it would
be better to have a separate service that did not include the
reliability mechanism to start with.

A few years ago I described my memory to Danny to see if he remembered
and could help me reconstruct my memory more completely, but he did
not recall it.

                                                        -- Steve