On Sat, 12 Jul 2008, Brett Monroe wrote:

> Honestly I would have expected similar results between the two 
> arrays. . . for a small to mid-size work load.  I would expect the 
> 6140 to shine as IOPS (OLTP) increased or streams are added due to 
> the increased cache size.  Also I imaging that a 6140 will scale 
> (IOPS and throughput) much better then a 2540 as you can attach more 
> trays to it.

IOPS is primarily dependent on the type of disks used and the number 
of spindles involved.  A larger cache does not help at all unless it 
would otherwise become full.  The cache only has to be big enough in 
order to ensure that there is always work queued for all of the disks 
and to deal with intermittent bursty writes.

A larger cache will help if there are several bursty writers 
which would otherwise overflow the cache and cause a writer to block.

The performance difference between these arrays is likely to be in the 
area of sequential I/O as long as the disks are of a similar type and 
performance.  This is when speed of the cache, interconnects, 
backplanes, and firmware, becomes most evident.

Benchmark data is formally posted so you can see for yourself what the 
difference is.

Bob

_______________________________________________
storage-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/storage-discuss

Reply via email to