Feature

Grid gobbledygook hits storage

Fearful as always of missing the boat, storage vendors are jumping on the grid bandwagon, unveiling visions that are about as useful today as a chocolate teapot.

Published: 15 Dec 2004
Grid computing might be the next big thing to hit IT, but without a sensible discussion of what grid computing actually means to storage, no amount of visionary talk, or blueprint designs or grid-enabled architectures is going to make a blind bit of difference to anyone.

Ultimately, grid computing will mean questioning the validity of the storage industry as a separate entity altogether.  
,
The latest culprit to spill grid goobledygook is EMC, in partnership with Dell, Oracle and Intel. This gang is building "megagrids" because apparently regular grids just aren't big enough anymore. Megagrids, according to an EMC spokesperson, are about applying resources in a network to a single problem at the same time.

Last time I checked, this was called clustering.

Network Appliance is saying the same thing. At a two-day event in New York City recently, NetApp briefed analysts and journalists on its "grid vision." Under the guise of "grid-enabling" its technology, NetApp is re-architecting its file system to support clustering and load balancing. That's it folks. So what is grid computing then?

Here are a couple of examples: The LHC computing grid links large clusters and storage systems in 87 computer centers around the world for the benefit of particle physicists. Another example is TeraGrid, an American effort to link nine large supercomputing centers for scientific use.

These projects offer a glimpse of grid computing, which provides a single global, standards-based computing grid that will do for data processing what the World Wide Web did for online publishing.

The storage I/O for these kinds of projects is off the scale compared with what most storage systems can provide today.

Grid-based storage throughput will be measured in hundreds of gigabytes to terabytes per second instead of megabytes per second. Aggregate storage I/O will be measured by the hundreds of millions to many billions of IOPS. Additionally, the storage must be capable of being geographically distributed while maintaining a single image. Finally, that storage will be capable of scaling both capacity and performance linearly up to tens of petabytes, exabytes and even yottabytes in a single image. (A yottabyte is a billion petabytes.)

Ultimately, grid computing will mean questioning the validity of the storage industry as a separate entity altogether. A senior analyst at Gartner recently pointed out that once the computer infrastructure becomes completely virtualized, the lines between networking, processor power, capacity, etc. will disappear. But that's probably a decade away, he says. In the meantime, we could keep the discussion of building megagrids to a minimum and try to understand what's really going on here.

Editor's note: Views expressed here are not necessarily those of SearchStorage.com.

Read more on Integration software and middleware