I think that there is also a general consensus that billing for CPU time is very difficult to do these days. Most programs are not CPU bound and CPU is plentiful. There isn't much time used. To add to this, a lot of processing is done in the kernel in such a way that it is hard to bill for it. Finally, bandwidth costs usually dwarf clustering costs. I'm sure there are a lot of high-performance people that might consider it, but for the high-availability folks, it's not much of a sell. On Dec 26, 2006, at 12:44 PM, Chetoo Valux wrote: Dear all, -- Jayson Vantuyl Systems Architect Engine Yard |
-- Linux-cluster mailing list Linux-cluster@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/linux-cluster