X
Innovation

A grid by any other name (eg: utility) is just a grid

I'm at GlobusWorld in Boston, where a relatively small herd of grid lovers has gathered to get an update on the state-of-the-state of the fundamentally disruptive technology (disruptive to most existing compute paradigms). I took meetings with the Enterprise Grid Alliance, MCNC (its grid-enabled North Carolina-wide backbone network is one of the more advanced of its kind in the world) and an outfit with some way-cool "hive computing" technology called Tsunami Research.
Written by David Berlind, Inactive
I'm at GlobusWorld in Boston, where a relatively small herd of grid lovers has gathered to get an update on the state-of-the-state of the fundamentally disruptive technology (disruptive to most existing compute paradigms). I took meetings with the Enterprise Grid Alliance, MCNC (its grid-enabled North Carolina-wide backbone network is one of the more advanced of its kind in the world) and an outfit with some way-cool "hive computing" technology called Tsunami Research. But my favorite bit of FUD-busting of the day came from leading grid researcher Ian Foster, who said that computing concepts such as Web services, utility computing, virtualization, data center automation, and the adaptive enterprise are all solving the same problem and should be solving it in the same way.
One of the event's major themes is how and why grids should be (and are) converging around interoperable standards. The Globus Alliance provides an open source-based toolkit (GT4) that no attempt at grid deployment should be without. Judging by the way Web services standards such as WS-Security, SOAP-based messaging and the WS-I's basic Web services profile are an integral part of the standard grid playbook, grids are rapidly turning into a proving ground for service-oriented architectures (SOAs).
With standards playing such a prevalent role in grid evolution, this week's event reminds me of the early days of Interop (now called Networld+Interop), when people -- mostly the innovators in bridging and routing technologies -- gathered to discuss and feature interoperable networking technologies well before networking exploded onto the scene to the point that it just became baked into the fabric of everything we do. Grid technology, as far as I can tell, is destined for the same future as networking was back in its embryonic days. As soon as the technology -- largely relegated to the academic and scientific communities for now -- has a few really big corporate success stories (note Sun's efforts with it utility computing rollout) under its belt and it's clear there's money to be made, the commercial feeding frenzy will begin and events like GlobusWorld will transition into much bigger events and showcases. Noticeably absent from the event is Microsoft. I'll have more tomorrow after I've had the opportunity to digest today's notes.
Editorial standards