Gridding My Teeth

Grid computing seems to be another case of too much vendor push, and little-to-no user pull

August 19, 2006

1 Min Read
Network Computing logo

1:15 PM -- IBM's at it again, pumping up the power of grid computing to handle enterprise requirements and applications. (See IBM Supercharges Grid.)

I'm not buying it. And apparently neither are a lot of enterprise customers. (See Enterprises Still Not Sold on Grid.)

Universities and government laboratories have led the charge on grid uptake -- and with good reason: They've got nearly unfathomable computational processing requirements, and budgets that don't necessarily grow to accommodate expansion or new opportunities.

Okay, so we know one big car maker is using grid for crash test measurements. (See Daimler Maps Grid Savings.) But if Boeing or Bechtel need more MIPS and gigaflops, they're probably just going to ramp up and buy more servers and processors. As my accountant once growled at me, "Leasing is an expensive way to buy."

Roy Rabey, an IT manager at gaming company Ensemble Studios, says as much in our story this week (noted above). Grid's not worth the premium, he told Byte and Switch. Grid also doesn't come with any prioritization schemes or service level guarantees -- at least not yet.Bottom line: I doubt that most enterprises are doing anything as compute-intensive as genome sequencing or long-term weather modeling. Just pushing payroll and inventory information where it needs to go would be a gracious plenty for most departments.

Think this is so much over-simplification? Do you believe enterprises can use grid to tap into major savings or productivity gains? Do me a favor -- don't email me, but hit the "Discuss" link below and let's find out where those of you in the trenches really are with your grid thinking.

Terry Sweeney, Editor in Chief, Byte and Switch

Stay informed! Sign up to get expert advice and insight delivered direct to your inbox
More Insights