What You Can Learn From HPC

Chances are, if you work in an ordinary, run-of-the-mill data center, or even the storage arm of a large enterprise installation, you're not going to need 1 petaflop of sustained performance, like the Roadrunner system IBM has supplied to the U.S. Department of Energy.

You probably also won't require a 1.1-Pbyte acceleration appliance for your filing systems, as NetApp announced today.

And chances are you won't be looking at 40-Gbit/s InfiniBand connectivity anytime soon, even though Mellanox unveiled the building blocks yesterday.

Even if you don't need any of this stuff, it's fun to watch someone who does. After all, there are lessons in high-performance computing for ordinary environments. Here are just a few of the issues faced by HPC that apply universally:

Clearly, there's value in seeing something work on a grand scale. Products and services that can work for big banks are likely to work for SMBs, eventually. Of course, there's an art to remodeling enterprise wares for the small shop, instead of simply selling dumbed-down or half-functional products. But that's another story.

Have a comment on this story? Please click "Discuss" below. If you'd like to contact Byte and Switch's editors directly, send us a message.