[e2e] Fwd: Camel's nose in the tent
day at std.com
Fri Aug 10 20:08:38 PDT 2001
At 10:53 PM -0400 8/10/01, David P. Reed wrote:
>At 07:29 PM 8/10/01 -0400, John Day wrote:
>>Actually, my understanding of SMTP is that it is intended to be as
>>much an application relay as X.400? No. I would agree. The ends
>>were the ends of a transport layer connection, i.e. TCP.
>The end-to-end argument applies at each protocol layer, not just the
>transport layer. Both the rationale for it, and the explicit
>description we wrote in our paper apply at application layers as
I would agree with that. Although I have been told by experts that
that isn't the case. But just for the record, does the application
layer have the same scope as the layer below it?
>You could claim that an SMTP connection is only one hop. But each
>relay step is specified in the protocol, and the UA's at source and
>destination are the ends.
Right, and to have reliable transfer you need an end-to-end protocol
that you don't have. So exactly how is the SMTP end-to-end?
>In regard to reflectors and listserv's, they are ends as far as SMTP
>is concerned - because they are UA's as far as SMTP is concerned.
And also, how does end-to-end apply to the data link layer?
More information about the end2end-interest