Fortunately the disk storage pools on TSM have a larger block size which
is much more suited to throughput.

Again the RAID type and read cache available will have an effect on this.
But as Allen says, you will see the performance ceiling much quicker on
SATA than FC, as random IO will have to come directly from the spindle(s).


Ian Smith
-------------------------------------------------------
Core Engineering - Storage




"Allen S. Rout" <asr AT UFL DOT EDU>
Sent by: "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
12/07/2007 15:19
Please respond to
"ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>


To
ADSM-L AT VM.MARIST DOT EDU
cc

Subject
Re: [ADSM-L] SATA disk?






>> On Thu, 12 Jul 2007 07:32:27 -0500, Andy Huebner
<Andy.Huebner AT ALCONLABS DOT COM> said:

> To me it is not the incoming, but the outgoing that is the greatest
> concern.  If the next pool is physical tape, be sure you have enough
> speed to keep up with your tape drives.  If the disks cannot keep up
> with the physical tape drives then performance will not be acceptable.
> Since many VTLs use ATA disks we know it can be done.


This is a really good point.  Keeping the number of read streams under
control will make a huge difference in your performance.

When I drain a SATA pool with a single thread, I can drive one of my
3592s at 80-90 MB/s.  When I put three of them on the job, aggregate
throughput drops to ~30 MB/s (yes, each drive getting ~10MB/s).

- Allen S. Rout



---

This e-mail may contain confidential and/or privileged information. If you are 
not the intended recipient (or have received this e-mail in error) please 
notify the sender immediately and delete this e-mail. Any unauthorized copying, 
disclosure or distribution of the material in this e-mail is strictly forbidden.

Please refer to http://www.db.com/en/content/eu_disclosures.htm for additional 
EU corporate and regulatory disclosures.