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

["Stephen J. Turnbull" <stephen@example.com>]

Curt Sampson writes:

 > 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.

If you had asked me, I would have bet on Google using cheap components
with lots of redundancy, because their business model needs to have
lots of redundancy anyway.  My guess is that anybody asking about the
difference between software and hardware RAID on TLUG is not a
candidate for comparison to Google.

That doesn't mean that a cluster or a multibus server or a multi-ATAPI
server isn't the answer, just that "that's how Google does it" isn't
relevant.

 >> 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.

While I'm not an expert, the computation I was demoing is a no-brainer
to parallelize.  My point is that the off-the-shelf solution I used
(PVM) performed a factor of 1000 worse than "sucky", because it didn't
work and play nicely with the OS.  The same thing is going to be a
worry for any distributed computation system.

FWIW, here's a description of the problem (written for my colleagues,
not for geeks, so grit your teeth) and links to the code:

http://turnbull.sk.tsukuba.ac.jp/Blogs/Software/ParallelVirtualMachine