If you take a big swig of vendor Kool-Aid from IBM Corp., Hewlett-Packard Co., Sun Microsystems Inc. or Oracle Corp. these days, it’ll taste a whole lot like grid computing. The actual flavour might be labelled (or mislabelled) as something else: utility, on-demand, autonomic, adaptive or even pervasive computing. But the upshot will be this notion of making miraculously cost-effective use of idle networked computing resources.
If you believe the various descriptions of grid (everybody has a favourite), we’re heading toward an IT nirvana where processing power and applications will be dynamically reconfigured and delivered from one big virtual pool of resources. Data will be secure but available from anywhere, and network complexity will be tucked out of sight.
It’ll be so great when it gets here.
Despite the full-court press from vendors, skepticism prevails among business and commercial users, including some senior IT executives on a panel I moderated recently. We were talking about where technology is heading, comparing practical realities to market spin, and I asked their opinions about what is today’s most hyped technology, fully expecting it to be Web services. “Grid computing,” everybody said.
What seemed to irritate my panelists most was the answer-to-your-prayers urgency of vendors with a few new products to push.
“Marketers are exploiting the perception that grid is an advanced technology and in many cases are applying the term to offerings that, at best, have only a tenuous relationship with the strict definition of a grid,” wrote Gartner Inc. analysts Carl Claunch and Anne Powell earlier this year. The analysts cautioned that grid computing was “well on its way to the Peak of Inflated Expectations” (I swear, I’m not making that up) in the Gartner “hype cycle model.”
Indeed, the grid projects making headlines tend to be group efforts involving universities, science or engineering companies and hefty federal grants. So when Oracle Chief Executive Officer Larry Ellison claimed a few weeks ago that half of his current database customers will be grid users in five years, he was just, well, being Larry.
Despite some compelling projects, grid computing is many years away from everyday use in business. That’s the conclusion of the story “Grids Extend Reach”, which examines the current state of this still-emerging technology.
What’s keeping enterprise users at arm’s length from grid computing? Quite a lot. For starters, the mystery pricing schemes, lack of management tools and fluid arrival dates for many of the announced products make it difficult to embrace. Few business applications were written with parallel processing in mind, which means major code rewrites for the rest of them.
The proprietary impulses of vendors also threaten to slow the adoption of grid computing, since open standards are so crucial to its operation. The first warning sign came last month, when Oracle announced that it was forming a consortium for developing commercial grid applications. Yet the Global Grid Forum, established in 1999, already has hundreds of members. How nicely Oracle plays with the Forum bears close watching.
The fact that grid computing is continuing to grow in its traditional strongholds — drug and biotech companies, universities and federally funded consortia — bodes well for it to mature eventually into serious business offerings.
But if you’re inclined to pass on the vendor Kool-Aid for now, you’re not missing much.