all results?
Why are both arrays not in each of the results? Also the VNX was all flash configured I bet those HDD's were just the vault pack disks and not part of the disk pool.
You wait for a bus forever, then two come along at once. So it is with EMC and SPC benchmarks, and all of a sudden we have both SPC-1 VNX8000 and SPC-2 VMAX 400K. The Storage Performance Council lists industry-standard random block IO performance (SPC-1) and streaming storage performance benchmark results (SPC-2). Each one …
That's correct on the vnx vault pack only on spinning disk so not used in the test.
They turned on a special caching feature which isn't available to customers without a waver from product management. The same turns off pretty much all data services in favour of more available cache. Purposely removed hot spares and looks like they played around with multipathing settings, whilst only testing against half of the available capacity after mirroring overhead.
So using 40 x 100GB SSD's resulted in something like 1TB of working set data to 256GB of Dram cache (4:1). Hardly real world, but gotta hand it to EMC they still know how to game a benchmark and even better don't seem to mind who knows it :-)
Haven't checked yet but if expect similar silliness on the VMAX.
Most vendors use various silliness to maximize their results in these tests. Similar to SPEC20xx CPU tests where the compiler guys have broken some of the codes using very narrowly targeted "optimizations" that happen to have a major benefit in some of the tests while being applicable to almost no real world code (not to mention that all results, even base, use PGO which is rarely used in real world code)
Not sure why EMC finally provided results other than that they didn't want to unless they could produce some results that make them look good. I wonder if the lack of them was hurting them in some RFPs, or they felt doing well will increase their sales? Since SSDs appeared on the scene array performance varies so widely based on the configuration you choose the results in these tests are not useful in any way.
EMC finally publish some results. For me they can rig them however they want, as long as it is disclosed. I am not an EMC customer or expert by any stretch but I did find the information on "Extended Cache enabler" interesting. Though I have to wonder what is the point of doing that to a VNX, discarding all data services for the best performance.
Does that mean the VNX is faster than XtremeIO or cheaper than XtremeIO? Because if data services don't matter and you want EMC, I would expect XtremeIO would be the product to look to.
At the same time I find it curious they did not run the VMAX through SPC-1.
Interesting to me in any case.
This post has been deleted by its author