Path: utzoo!mnetor!uunet!lll-winken!lll-tis!ames!ll-xn!mit-eddie!uw-beaver!tektronix!decvax!dartvax!eleazar!ericb
From: ericb@eleazar.Dartmouth.EDU (Eric J. Bivona)
Newsgroups: comp.protocols.appletalk
Subject: ATP and flow-control
Message-ID: <8763@eleazar.Dartmouth.EDU>
Date: 27 Apr 88 00:03:36 GMT
Reply-To: ericb@eleazar.Dartmouth.EDU (Eric J. Bivona)
Organization: Dartmouth College, Hanover, NH
Lines: 27
Keywords: transaction based protocols, flow control, congestion
Summary: looking for referneces


I'm curious about work that's being done to make transaction based
protocols congestion-sensitive, perhaps dealing with rate based
flow control.  I'm interested in general, and about ATP in particular.

I'm curious to know if anybody can give pointers to research and/or
literature on this, and (in a real life scenario) what people think
about Apple's scheme to specify time outs.  I know it's easier to
implement static timeouts, but it's just that: STATIC.  It doesn't
respond well to congestion (it doesn't respond) and certainly adds
to it if you blow an ATP timer.

I've heard that Industrial Mac's have a rewritten network package that
does do dynamic flow control in ATP, anybody know any details?

In terms of general research I know about the V kernel at Stanford,
and I'm trying to get a hold of Cheriton's paper on VMTP from
SIGCOMM '86.  I also know of the RFC for NETBLT, but I don't have
a copy of this yet.

Thanks!

-Eric
-- 
 eric.bivona@dartmouth.edu

"Forgive me Father, I want to sin"
"I don't do funerals."				Doonesbury