- casino online slots

Snap Server 550 Screams

by Administrator on March 5, 2006

Okay, we talked about common sense testing a lot. Here is a test report we just completed on the Snap Server 550 at TPI Technologies.

First Generation Test Lab Report

As you can see, the test is heavy on “black box” performance measurement using Iometer and IOzone. It begins to get to application-specific tests — though, this time, we elected to use published test methods rather than our own workload generator. I added qualitative evaluation as well.

Jeremy Evans, CEO of TPI Technologies, should get the kudos for the testing that was done. I think he did a bang-up job. He managed to get straightforward comparisons made of Snap 550 and several competitors in terms of basic IOPS and throughput. Going forward, I want to see more comparisons under workload, plus more MSRP data included.

Disclosures abound in the report. Jeremy’s (and the lab’s) test time was covered by the vendor. The reporting of results received, however, are outside the vendor’s control. Adaptec/Snap said they had a hot box. We didn’t fully believe their claims, since it runs a modified Linux OS, which typically suggests slower performance. They took the challenge and the box lived up to their claims.

When you read this report, assuming you do, I invite you to provide constructive input: what else could we add to the format that would make it more useful to you?

Like Adaptec/Snap, we are putting our money where our mouths are…

{ 3 comments… read them below or add one }

alan s March 5, 2006 at 4:01 pm

Hmmm, the link to document seems to be broken.

Administrator March 5, 2006 at 4:07 pm

I posted an interim, forgetting that the final was on my laptop. The final has been posted and the link now works.

Sorry for the momentary lapse of reason.

ed_zachary March 9, 2006 at 8:05 pm


Why no NetBench results, since so many of the players seem to be using that as the standard with VeriTest?

I like the straightforward testing but I would like to see some other benchmarks, like SPECsfs or IOmeter. The debate about their accuracy aside, it would seem to help compare apples and apples.


Previous post:

Next post: