Infiniband (was RE: Beowulf: A theorical approach)
billm at troikanetworks.com
Thu Jun 22 12:09:12 PDT 2000
Greg, Tony -
Looking at the infiniband info
it is intended to be a 'network approach to IO', and is
limited to more-or-less local access (<1000 meters). As
such it is oriented to clustering of systems - not just
server area networks or storage area networks. One of the
goals, apparently, is to have VI Architecture implimentation
on top of Infiniband - which is a fast, efficient cluster
From: Greg Lindahl [mailto:glindahl at hpti.com]
Sent: Thursday, June 22, 2000 11:26 AM
To: Tony Skjellum
Cc: beowulf at beowulf.org
Subject: RE: Infiband (was RE: Beowulf: A theorical approach)
> I will say that several people have mentioned that Infiniband is for
> server area networks, not system area networks (ie clusters).
This is true, if the rumors I've heard are true. And server networks are for
accessing storage, with few conversations and huge block sizes, not for tiny
messages to any of thousands of hosts in the entire machine. There will be
an annoying limit on the # of connections, and the standard only guarantees
1 outstanding non-connection message.
The one thing that Infiniband will do is provide a much better bus than PCI.
PCI-X is better in ways other than just large-transfer bandwidth, and
Infiniband is still better. I have yet to see any sign that "native"
Infiniband switches are going to be good. IBM's announcement of 8-way 6
megabit switches (perhaps I've got the details wrong, doesn't really matter)
available 12 months from now just doesn't excite me much.
So what will we do? We'll stick Myrinet cards (or GigabitN Ethernet or
Quadrics or whatever) on Infiniband just like we stick Myrinet on PCI today.
Beowulf mailing list
Beowulf at beowulf.org
More information about the Beowulf