Path: utzoo!attcan!uunet!decwrl!sgi!vjs@rhyolite.wpd.sgi.com
From: vjs@rhyolite.wpd.sgi.com (Vernon Schryver)
Newsgroups: comp.protocols.tcp-ip
Subject: Re: Parallel network problem
Message-ID: <63581@sgi.sgi.com>
Date: 7 Jul 90 22:46:53 GMT
References: <5A07060B131D020F-MTAABB*John.Curran@VEC.IMS.ABB.COM>
Sender: vjs@rhyolite.wpd.sgi.com
Organization: Silicon Graphics, Inc., Mountain View, CA
Lines: 29

In article <5A07060B131D020F-MTAABB*John.Curran@VEC.IMS.ABB.COM>, John.Curran@VEC.IMS.ABB.COM writes:
> 
> ...  Is load-balancing actually a lower-layer activity?


That would seem to depend on whom you ask.

Notice that many people in X3T9.5 think all of network management belongs
in the link layer, including all necessary tools such as reliable circuits
that can pass tens of thousands of bytes between management agents.  (Yes,
there are those in X3T.5, the ANSI committee doing FDDI, who have been
publically very much not dismayed by the prospect of SMT (FDDI Station
MangemenT) information exceding the FDDI frame size.)

There are others who think that SMT is overly complicated, that none of the
optional frames will ever be implemented by enough vendors to be usable
even by those who favor link-layer network management, that the absense of
authentication and authorization for remote actions dooms any hope of using
remote acting SMT frames in any customer site where users are not trusted
(e.g. any university or large company) and so at almost any site that
really needs remote FDDI ring management, and that the last change that made
SRF's mandatory was unfortunate.

I think load-balancing belongs above the link layer somewhere closer to
the machinery that bangs windows and such, but then I'm contrary.


Vernon Schryver
vjs@sgi.com