The Olympics of Storage
by Claus Mikkelsen on Oct 2, 2007
As much as I would like, I will not win a gold medal in next years’ Olympics in Beijing. Having been a sprinter in an earlier part of my life, I will have to hand over that glory next year to Tyson Gay, who has been pretty impressive of late in the 100 meters. Not that I wouldn’t enjoy the excitement, fame, and fortune that would come as a result of standing on the winners’ platform listening to my national anthem, I just know it cannot happen. In other words, I know when not to compete. This isn’t an “odds” game. What’s the message here? If you’re competitive and believe there is a million to one shot of winning, you will definitely take up the challenge. If there is absolutely no way you can win, you will choose to not compete. More on this in a bit…
Yesterday, HDS announced some pretty impressive SPC-1 results with our USP V storage platform. In fact, we achieved record breaking results which is no surprise to us. The SPC (Storage Performance Council) provides independent performance testing of workloads most relevant to the “real life” applications used by most of our collective customers. No, it is not a perfect predictor of what any customer might achieve in performance, but it is a pretty darn good indication of what to expect when comparing the performance from various vendors. A lot of very smart storage people contributed to the metrics measured by the SPC, and it is not to be taken lightly.
So, we did achieve phenomenal results, and yes, we’re proud of that. Almost twice the IOPS of IBM’s high-end array, for example, but what is interesting to me is not that, but how this would compare to what isn’t there – EMC. The SPC is a very legitimate performance organization that is supported by all of the major storage vendors. The SPC is supported by Hitachi Data Systems, IBM, Sun, HP, Dell, Fujitsu, 3Par, LSI, and others. That is, the SPC is supported by every major storage vendor and almost all of the smaller players. But who is missing? Yes, EMC, who originally joined the SPC and then bailed.
And why did EMC bail from the SPC? Let’s get back to the first paragraph. When you know you absolutely cannot win a contest, you’ll choose not to participate and address the deficiency with a multitude of excuses as Chuck Hollis babbled in his blog on July 12, 2007. The fact that EMC does not have a high-end storage product that can compare with the performance of HDS (and even IBM) is quite evident.
But Chuck didn’t stop making excuses in July. In his most recent blog he said “We’ve been in a lot of face-off benchmarks against the other guys. Sometimes we win, sometimes they win. A few of the vendors never seem to win (usually the do-it-in-the-storage guys), but that’s another story”.
So what’s the other story? As many of our customers have told me, IBM is significantly faster than EMC and HDS is significantly faster than IBM. I think this is a good lesson. Although screaming performance is not important to everyone, lack of performance certainly is and all of us, EMC included, spends a lot of R&D dollars on something that Chuck implies is unimportant.
As InfoStor put it: “However, EMC does not make its performance results public, nor does the company participate in industry standard benchmarking tests offered by organizations such as the Storage Performance Council, a practice that is commonplace with most of EMC’s competitors.”
And as Evaluator Group said, the invitation stands:
Chuck Standerfer called the results “very impressive,” and encouraged other enterprise storage vendors — namely, EMC — “to graciously accept this challenge and publish their SPC results so that customers would have a valid independent source for comparisons.”
So check our results in detail on the SPC Website, and the next time you’re thinking of a storage purchase, consider why one of the big guys in storage has chosen not to compete. We know why.
Let the games begin…