Re: [tlug] SATA software RAID or SAS hardware RAID?

[Curt Sampson <cjs@example.com>]

On Fri, 15 Dec 2006, Stephen J. Turnbull wrote:

> For me the killer fact is that people who depend on reliable servers
> for their bread and butter (eg, Akamai, IIRC) still are willing to pay
> for gold-plated SCSI/RAID solutions.  Do you think they're just being
> excessively conservative?

Yes. Google also depends for its bread and butter on reliable storage
of a very large amount of information, and they use cheap rackmount PCs
with cheap IDE drives. Others do the same.

> Also, while I'm no expert, getting multiple anything to perform better
> than a single piece of hardware seems to be non-trivial.

Well, it depends on what exactly you're trying to make it perform. The
cost of making something perform reliably can be altered dramatically
by changes in the structure of the system and the application. So
Google, for example, manages to avoid the quite difficult job of keeping
a single machine and set of drives up and running despite hardware
and environment failures by having a system that doesn't care if a
particular machine vanishes in a puff of smoke.

Gregory Pfister's book _Clustering_ has a lot of good information on
this sort of thing. And also excellent discussions of the kind of issues
you encountered with your example of a program that ran more slowly on a
parallel system than on a single-threaded one.

cjs
--
Curt Sampson       <cjs@example.com>        +81 90 7737 2974