Girding for Grid

On paper, the grid computing concept makes such great sense -- computing and storage resources pooled across the wide area via high-speed links. In reality, though, it's hard to see how the concept can rise above its fairly niche-y use by academics and the research community.

That's not a vertical market to take lightly, given the fact this group had a big hand in shaping that mildly popular networking idea we know now as the Internet. What universities and R&D labs have in common are that they tend to be highly subsidized, and their user groups are often closed, finite and change little over time. That makes them about as distinct from enterprise users as they can be.

There's another issue with grid that's getting more attention as vendors talk up the concept: the limited number of applications. Complex web server transactions that require multiple database lookups and additional handling don't just happen -- turns out such ordinary enterprise applications would need a fair amount of re-engineering to work seamlessly across a computing grid.

Most enterprises won't bother with re-working existing applications or buying grid-enabled ones just for the bragging rights. That message is starting to get through, at least in part. IBM said this week it's working with a couple software companies to try to make grid more useful and, uhh, applicable to more users.

It's a start toward making grid something bigger and maybe less elite. Or maybe it will just expose grid's weaknesses or inappropriateness for broad enterprise use. Again, it's a cool concept that may be ahead of its time, or out of time completely. Write me and let me know whether you and your organization are girding for grid, or just waiting for the hyperbole express to run its course.Terry Sweeney, Editor in Chief, Byte and Switch