Dr Adam J Trickett <adam.trickett@???> wrote:
> I thought the evidence now suggested that the
> actual failure rate of a disk is proportional to
> it's age with only a small bath-tub effect and
> there was no longevity difference between modern
> SCSI or PATA/SATA disks at all.
I don't have statistically significant data to
draw on, but my personal experience involves lots
of SCSI drives that just keep running.
> In the early days SCSI disks were both more
> reliable and more flexible and potentially
> faster, now the only difference is that they
> cost a lot more per Mb.
There's still significant differences in the
drives.
Take Seagate for example, comparing the Barracuda
ST3320613AS 320GB SATA drive [1] with the Cheetah
ST3300007LC 300GB U320 SCSI drive [2].
Ignoring the obvious difference in spindle speed -
10k RPM for the SCSI, 7.2k RPM for the SATA - the
physical geometry is different. The SATA drive
achieves its capacity with one platter and two
heads, whilst the SCSI uses four platters and
eight heads. That suggests to me a much lower bit
density on the platter, which should result in
better signal-to-noise and less reliance on signal
processing to actually read or write bits.
The design of the SCSI drive is more conservative
than the SATA - the former is the largest capacity
drive in the family, the latter is the smallest -
which is what you might expect where reliability
is a major concern.
Nick.
[1]
http://www.seagate.com/docs/pdf/datasheet/disc/ds_barracuda_7200_11.pdf
[2]
http://www.seagate.com/docs/pdf/datasheet/disc/ds_cheetah10k.7.pdf
--
Nick Chalk ................. once a Radio Designer
Confidence is failing to understand the problem.
