going to the moon vs boiling the ocean

It seems to me that saying "The Open Grid Forum should commit all its available resources to the goal, that before this decade is out, commercial and academic organizations will build real operational grids using OGSA based components!" is like President Kennedy saying that we should go to the moon using multi-stage rockets that used JP-5 fuel and core memory guidance computers. Again, I'm not trying to knock OGSA, but I'm trying to set the right level for this discussion and the document. The goal statement of the organization should not include a specific technology solution. The terms "grid" and "grid computing" have a long, storied, and somewhat checkered past. The term has gotten quite muddy. Articles like http://weblog.infoworld.com/gridmeter/archives/2006/07/a_broader_scope.html draw this in sharp relief. We should take advantage of this inflection point to recapture the mindshare around "grid", by setting and stating goals crisply. I assert that grid computing is utility computing, on a range of hardware, in a range of architectures - it's not just HPC computing, stealing cycles from one another's supercomputers any more. (Nor is it, really, about cycle recovery from desktops or embarrassingly parallel applications like SETI@Home. Those are interesting, they got the term out into the marketplace of ideas, but we're not going to be developing organizational strategy around those.) Grid computing is about these things: - infrastructure virtualization - resource pooling & sharing - self monitoring & improvement - dynamic resource provisioning - highest quality of service OGF's goals and the TSC's strategic roadmap need to capture all this. Well, *I* think so. I'd be interested to know what others think. Best, chris
participants (1)
-
Chris Kantarjiev