Too many peers considered harmful?

For the benefit of those who are not close observers of the Australian
academic/research networking scene, we are busy putting a national
network back together again.

In this process, we are looking to coordinate a number (approx 20) of
largish caches (from 15-30GB currently) with quite a large number of
clients driving them, into a mesh which eventually minimises use of
international links - translating directly, btw, into dollars saved.
Volume matters, request hit rates don't.

In arriving at an initial configuration, we would really like to hear
from anyone who is attempting this without informing the rest of us
just yet. In particular, we would be interested in opinions, assertions,
or hard evidence about how things scale as one adds big chunks of
unsatisfied demand to a unicast or multicast mesh and whether ICP
traffic (at a lower charged rate) can overwhelm saved traffic charges
elsewhere. (thus the provocative subject heading above)

Brian Denehy
(on behalf of the AARNet2 cache working party)
-----
Brian Denehy, Internet: B-Denehy@adfa.oz.au
IT Services MHSnet: B-Denehy@cc.adfa.oz.au
Australian Defence Force Academy UUCP:!uunet!munnari.oz.au!cc.adfa.oz.au!bvd
Northcott Dr. Campbell ACT Australia 2600 +61 6 268 8141 +61 6 268 8150 (Fax)
Received on Fri May 16 1997 - 08:08:46 MDT