www.aspnews.com/analysis/article.php/1576041

Back to Article

Weekly Review: It's a Tangled Grid Hype Weaves By Phil Wainewright January 28, 2003 Every year, a few emerging concepts capture the imagination of marketing people in the technology industry and consequently get twisted out of all recognition. This year, it looks like it will be the fate of grid computing to be abused in this fashion. Read and React "Grid computing is not as tightly integrated as a cluster, but nor is it as loosely coupled as a network. It's somewhere in between the two. Alternatively, if you're in marketing, it's anything between the two." Give us your feedback in the ASPnews Discussion Forum Grid's turn has come about because the technology has just reached that level of maturity where it's more-or-less usable in certain, very limited, applications. This creates practical case studies that the technology's supporters can now cite as evidence that all their grand predictions for its wider potential are about to be realized. Such claims, of course, are usually made without any examination of whether or not the case studies in question actually have any relevance for more mainstream applications. That would be to ignore the real issue here: grid is ready to be hyped. Grid Beyond the Hype For those of you who haven't yet heard the hype — as well as for those of you who have, but who've been left feeling none the wiser for it — perhaps I should briefly explain the concept of grid computing. It's what happens when you share a computing task across several separate computers, with each of them contributing any suitable resources they have available at the time. It's not as tightly integrated as a cluster, where all the computers are uniquely dedicated to the task in hand, but nor is it as loosely coupled as a network, where each of the computers runs only their own individual applications. It's somewhere in between the two. Alternatively, if you're in marketing, it's anything between the two — and anything else remotely related. Hence IBM's announcement this week (see IBM Launches Commercial Grid Offerings). Its ten grid computing packages are a marketing triumph — but then what else would you expect from the company that, according to its recent worldwide ad campaign, says buying its services is the closest you'll ever get to owning a Universal Business Adapter? IBM's packages span the gamut of grid computing, starting just short of clustering and stopping close to plain old wide-area networking. There's an analytics acceleration grid, used in financial services or life sciences, and a closely related engineering design grid. Harnessing closely linked machines within a local area network environment, there's little to distinguish these offerings from a cluster. There's an information access grid, available in versions for life sciences and government, which seems oddly similar to what others might describe as a business intelligence or content management solution, with its ability to provide "unified data access [to] ... nonstandard data formats." Moving out to the wide area network, a design collaboration grid enables "data sharing and distributed work flow across partners" — just like the Internet. Most curious of all is the IT Optimization Grid, which claims to "help customers exploit ... underutilized compute and storage resources." The idea of turning idle and obsolete equipment back into useful assets with a light sprinkling of grid computing sounds rather like a medieval alchemist promising to turn base lead into precious gold. I'm also surprised to see IBM targetting this package at the financial services industry. If the sector does have plants lying idle, this may not be the best time to draw attention to the fact — not so much a case of understanding the customer's pain as rubbing salt into a fresh wound. On the one hand, IBM's marketing people should be congratulated at packaging this hitherto somewhat arcane technology into a form that makes sense in a business environment. But on the other, lumping this broad range of these solutions under the same umbrella will do nothing to allay confusion about what grid computing actually is. Globus Project Makes Progress This is unfortunate because, despite the hype, significant progress was made earlier this month with the launch of a version 3.0 toolkit by the the open-source Globus Project, which has been the prime mover behind the establishment of industry-wide standards for grid computing (see coverage at our sister publication, GridComputingPlanet.com, Next-Generation Globus Toolkit Released). Once the hype has died down, there's no doubt that grid computing is set to make a significant contribution to information technology in the 21st century. But for the next year or so, vendors will be leaping on the grid bandwagon with whatever products and services they can plausibly find a place for on board. Grid's moment of hype has arrived, which means the real story is going to have to take a back seat for a while. Do you have a comment or question about this article or the ASP industry in general? Speak out in the ASP Discussion Forum.