Grid architectures and middleware

Grid architectures and middleware

Distributed and massively multi-processor computational Grid services may involve synchronous hosting on distributed computing resources, including adaptation to the processing capability of the different hosts, with rapidly scalable computation, communication and storage solutions. They may include the management of asynchronous massive, highly distributed, intermittently connected, voluntary networks, incorporating peer-to-peer cooperative working, with remote resource allocation and job management. Users need a single transparent interface to different job schedulers, all with a high degree of portability.

These are used, to differing degrees, in predicting human-induced climate changes over a fifty-year period, in simulations of environmental effects, in visualisation of complex and massive data arrays and in medical diagnosis involving huge image sources.