On 21 Feb 2003 at 11:59, Jakab Laszlo wrote:
> Unfortunatelly the hardware budget should be keept as low as possible.
> I was thinking is there could be reliable solution based on dual processor
> and ATA 133 raid mirroring normally with some gigs of memory.
Gigs of memory are not as much important as much badnwidth you have. For these
kind of databases, a gig or two would not make as much difference as much
faster disks would do.
If you are hell bent on budget, I suggest you write a custom layer that
consolidates results of query from two boxes and throw two intel boxes at it.
Essentially partitioning the data.
If your queries are simple enough to split and consolidate, try it. It might
give you best performance..
Bye
Shridhar
--
Blutarsky's Axiom: Nothing is impossible for the man who will not listen to
reason.