>>>>> "M" == M C Srivas <[EMAIL PROTECTED]> writes:
M> it is: one of their conclusions were that overall system
M> utilization was probably better if they divided that 8-10 Gig
M> across the several Macintoshes so that each Mac had about 15-20M
M> of disk cache. Something about one person's working-set purging
M> all the others'. I think the paper appeared either in Usenix 1992
M> or in one of the AFSUGs proceedings.
I use AFS for distributing binaries and the like for applications that
change frequently, mainly in-house applications. This stuff stays
locked down pretty much, since there is always someone using it. I
dont have user data, or any read-write stuff for that matter, in
AFS. Atypical perhaps, but I'd be surprised if the findings you cite
hold true in this scenario.
If anyone has a citation for this paper (honey?) I'd love to read it.
M> I would rather do this by putting a volume replica across the WAN
M> and setting appropriate fileserver preferences; however AFS does
M> limit one to 13 replicas. Not very useful for data that changes
M> more frequently, but then my guess is so would be any solution
M> that expects to maintain data-consistency across a WAN while
M> expecting data to change frequently.
Can't do this until servers can continue to serve Read-Only volumes
while partitioned from quorum. Waiting..... :)
-Rens
