Subject: Re: Serious longstanding problem with TCP
To: None <tech-net@NetBSD.ORG>
From: Rafal Boni <firstname.lastname@example.org>
Date: 11/22/1995 13:55:11
In message <199511221500.KAA25454@Collatz.McRCIM.McGill.EDU>, you write:
[...stuff re: slow-start, crappy WWW perf. and such...]
-> Perhaps slow-start needs to be rethought in view of WWW, which I think
-> is probably the first protocol to carry on a lot of TCP conversations
-> sequentially, sending only small amounts of data over each one.
Regardless of slow-start or not, you've still got the 3*RTT cost of
setting up a connection... I believe there is work being done on
T/TCP (x-action TCP) that allows you to skip the 3-way handshake in
some subset of cases...
-> Personally, I think a preferable solution would be to redo the WWW
-> protocol so it doesn't keep tearing down connections and creating new
-> ones, precisely because connection setup and destruction has high
-> overhead. Slow-start just aggravates this by increasing the effective
-> connection-setup overhead. (To be sure, slow-start has its benefits,
-> but they don't kick in in these cases.)
Well, T/TCP ought to allow you to get by without some of that cost,
but there is work being done on making the HTTP protocol slightly
more intelligent in its' assumptions (ie, keeping connections around
much longer, etc.)... Un-suprisingly enough, it's being called HTTPng.
I've got links to all this stuff somewhere, so if anyone is interested
drop me a message and I'll look for 'em.