Yep, we'll judge WHOLE DARN storage config. Problem with that?

Behold. The benchmarkers have brainstormed a fresh filer performance measure – SPEC SFS 2014. It looks at end-to-end filer performance instead of looking at component levels, as did the previous SPECsfs 2008 measure. SPEC SFS 2014 stands for the Standard Performance Evaluation Corporation and its Solution File Server 2014 test …

[Lusty]

Big Data

I would expect anyone smart enough to be implementing big data analytics to work out requirements for their storage anyway so a benchmark is sort of pointless. These kind of benchmarks are aimed at c-level staff looking at options for their storage which is a different process to how storage tends to be bought for analytics.

[Nate Amsden]

Re: Big Data

using big data analytics to work out storage array requirements? How high are you? I'd say give me whatever you are on but I'll hold off on drugs until the superbowl (go hawks)

[Lusty]

Re: Big Data

Not even close to what I said, how high are you?

[Anonymous Coward]

Re: Big Data

You're both sober !!!

What does the 'C' stand for in "SPEC"? This test is nothing but a campaign of donations to get your "guy" to the top of the list. If you happen to go with that "guy", you'll find out that performance is not as advertised by the campaign agency...but too late to change. NPO or not, somebody feeds the kitty.

[Lynrd]

Without benchmarks

there can be no benchmarketing

[Iozone]

Things to note

Nice photo. I'll let the readers figure out why I said that :-)

Things I might note:

* Platform and protocol independent (Windows, Unix, SMB[1..3], NFSv[1..4], Ext3 over iSCSI ...)

* User definable workloads. (Permit folks to model their production workloads)

* Market segment focused results.

* Easy for SPEC to add more standard workloads quickly.

* Flexible architecture that will be a foundation for the future. (other workloads and APIs)

* Support for complex workloads. ( Mixtures of data and meta-data )

* Built by folks that have do this for a living, and donated to an NPO so that it would be open & fair.

Thanks,

Iozone.org

P.S. In your article you have "up to data" should read "up to date" ( Yes, I sometimes do that myself )

[cphollis]

Better than a kick in the head

Hey, for what it's worth, this new benchmark is loads better than what we've had to work with in the past. Real-world application focus, architecture agnostic, etc. -- I'll take this vs. the previous anachronisms we've had to work with.