Alan Cox wrote:
Have you tested that? It's not clear to me that that should be so.
Physical characteristics determine the number of platters that can be in
a drive, and you generally see a manufacturer release several drives of
different capacity at the same time.
What makes the difference is the number of heads.
A 320GB array built of 8 x 40 GB disks has 8 sets of heads and can do a
lot more ops/second as a result. It can als have 8 outstanding commands
at a time unlike any pre NCQ 320GB disk.
Ah, there we differ. The number of heads can be important - we had one
of these at one time[1]:
http://www-03.ibm.com/ibm/history/exhibits/storage/storage_2305.html
As I recall, one model had two heads per track, the other had one.
"Seek" time was the time for electronic selection of the head, and the
model with two heads had half the latency. That would also be the one
that could write at 3 Mbytes/sec (less than four seconds for the drive's
capacity!).
However, the reason the number of heads is important is that it affects
the amount of data that can be read/written without seeking. Data
density also affects that.
The original question back then, was my boss's boss was wondering which
would be better for databases, the big (100 Mbyte) 3330s (also mention
around there) or the smaller (29), older 2314s. I don't know that the
question was resolved - except we did use the 3330s for the databases -
but I think the 3330s would have won because of other features - RPS,
the ability to connect to four channels (think wires to the computer).
However, if one adds a black box containing bunch of drives, a CPU and
cache management software, well it's not really a disk drive any more,
even of the host sees it as one.
[1] The obvious use for it was paging/swapping, but that is not what we
used it for. The I/O bottleneck was a database index.
--
Cheers
John
-- spambait
1aaaaaaa@xxxxxxxxxxxxxxxx Z1aaaaaaa@xxxxxxxxxxxxxxxx
-- Advice
http://webfoot.com/advice/email.top.php
http://www.catb.org/~esr/faqs/smart-questions.html
http://support.microsoft.com/kb/555375
Please do not reply off-list
--
fedora-list mailing list
fedora-list@xxxxxxxxxx
To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
