Beowulf: A theorical approach

[James Cownie]

>  The problem right now really isn't in link speeds (though better
> link speeds are good), its in how close/far the network interface is
> from the CPU.  COTS HW doesn't place a high value on direct access
> to IO devices - there is a higher value on a standardized bus
> interface to allow different system components to be integrated and
> updated independently.  A "supercomputer" can have the network
> engineered directly into the node architecture.  This is a huge
> advantage.  Luckily, this advantage has the most effect in only some
> programs.

If Infiniband does all that it is supposed to do, then it will rapidly
become the network of choice, since it _does_ have support for direct
(user-space) access to the comms, and has some nifty switches.

Of course in the short term it will be limited by the CPU side
interfaces being PCI, but that's only the same limitation as
for Quadrics, Myrinet, SCI and so on. 

Once it becomes the standard for connection to storage it _should_ be
cheap, and a standard component of any "server-class" commodity
machine, whether IA* or other architecture. (IBM announced that
they'll be selling their interface chips and switches just today).

So, I expect that Inifinband will be engineered intimately into the
node architecture of COTS hardware, and that will help a lot.

It'll be interesting how long it takes before the Linux drivers are
available !

-- Jim 

James Cownie	<jcownie at etnus.com>
Etnus, Inc.     +44 117 9071438
http://www.etnus.com