I have been working in the NAS space for eight years. During this time I’ve reviewed a lot of NFS benchmark results on the SPEC website and you can too. Results for the current version of the benchmark, called SPECsfs2008, can be found here. (Ok, at this point you’re thinking I’m the Biggest Loser. Sorry, you’ll have to go to my Facebook page for that.)
I like SPEC for two reasons. First, it does a nice job of emulating large-scale file server environments, at least that’s what I am told by customers. Second, the world of performance benchmarking is full of unsubstantiated claims, but with SPEC you can count on a fair, apples-to-apples comparison of NFS server products.
The one thing that has always bugged me about SPEC is the waste. I don’t know about you but my wife has four “waste” bins in our kitchen, one for bottle & can recycling, one for paper & cardboard recycling, one for biodegradable food for the compost heap, and one for true garbage. We know waste when we see it.
The waste I am referring to on the SPECsfs2008 site is the overuse of disk drives. If you take a look at the four highest SPECsfs2008_nfs.v3 ops/sec results on the SPEC website, you’ll find that Exanet used 592 disks to achieve 119,550 ops/sec with 2.07ms ORT (overall response time). Huawei Symantec used 960 disks to achieve 176,728 ops/sec with 1.67ms ORT. And NetApp used 324 disks to achieve 120,011 ops/sec with 1.95ms ORT. By contrast, Avere used only 79 drives to achieve 131,591 ops/sec with 1.38ms ORT. Doing a little math, Avere achieves 8.2, 9.0, and 4.5 times more ops/sec per disk used than the other vendors.
So now you’re asking…
Is this a new diet out of Pittsburgh? Can Avere help to slim-down my overweight NAS? How does Avere do it?
No. Yes. We separate performance scaling from capacity scaling and more efficiently deliver both. Check out the product pages on our website for more.
SPEC® and the benchmark name SPECsfs®2008 are registered trademarks of the Standard Performance Evaluation Corporation. Competitive benchmark results stated above reflect results published on www.spec.org as of Oct 12, 2009. Above we compare all SPECsfs2008_nfs.v3 results that achieved greater than 100k ops/sec throughput and any level of ORT. For the comparison we calculate ops/sec per disk by dividing the reported ops/sec throughput by the total number of disks used in the system under test. For the latest SPECsfs2008 benchmark results, visit http://www.spec.org/sfs2008.
