[TriLUG] SAS vs. SATA performance

[Kevin Flanagan]

I don't think that they were saying that the specs were what made the
difference, but rather the grade of drives that are commonly used.  The
difference between different Seagate disk platter sets could be something,
the interface surely is different but doesn't make the difference.  Baracuda
vs Cheeta drives, etc.

I have skimmed the Google MTBF studies, they sure make a case, but at least
I get 3 year warranty on SAS vs 1 year on SATA in Proliant servers.




On 7/26/07, Robert Dale <robdale at gmail.com> wrote:
>
> On 7/26/07, Kevin Flanagan <flanagannc at gmail.com> wrote:
> > There's a lot more to the difference, the drives themselves may in that
> > order, but consider the controllers, channels etc.  Seek times, buffers,
> > rotational speed are surely considerations, but SATA disks are generally
> on
> > a smaller number of channels and ports than SAS drives.
> >
> > Another consideration is the MTBF, HP advised me that the MTBF of a SAS
> disk
> > is 3 times greater than that of a SATA disk.
>
> I wonder if he gets paid 3 times for saying that...
>
> SAS and SATA are just buses, not a materials requirement, not a
> manufacturing process, and not a mechanical specification.  Google and
> CMU, separately, have put out some statistics on drive failures which
> show MTBF means absolutely nothing and drive longetivity is completely
> unpredictable.  Of course, the latest news is that drive failure is
> caused by 'magnetic wobbles'.  So unless these are non-magnetic SAS
> drives, you can safely ignore any MTBF sales-speak.
>
> HTH
>
> --
> Robert Dale
> --
> TriLUG mailing list        : http://www.trilug.org/mailman/listinfo/trilug
> TriLUG Organizational FAQ  : http://trilug.org/faq/
> TriLUG Member Services FAQ : http://members.trilug.org/services_faq/
>



-- 
+---------------------------------------------------+
Doing my part to piss off the religious right.