Grid Computing Gets Small

The US and Japan have successfully demonstrated one of grid computing´s long-standing holy grails - dynamic, on-demand provisioning of bandwidth and interoperability between high-performance resources in two national research testbeds. The automated interoperability between Japan´s G-lambda project and the US´s Enlightened Computing project was demonstrated 11 September at the annual Global LambdaGrid Workshop (http://news.ncsu.edu/releases/2006/sept/documents/global lowbargrid.pdf) in Tokyo. The demonstration featured some of the most advanced research facilities in both nations, highlighting new middleware capable of reliably coordinating both network and computational resources as well as other protocol and interface technologies. Advances in grid computing technology have tended to focus on large-scale research deployments like this one, but smaller deployments are beginning to get headlines as well. This shift could change the way we view this field-as long as grid architects are willing to expand their vision of a grid beyond raw network speed and CPU aggregation