Two-layered metadata service model in grid environment

Abstract

The primary problem of data management in grid environment is how to effectively organise the geographical distributed storage devices to support the collaborative data operations. This paper focuses on how to resolve the above problem and presents a two-layered metadata service model in a grid environment which uses the locality of users' distribution and provides a platform for grid data management. It also presents a metadata replica model, which uses a sparse strongly connected graph to connect replica metadata. In this way it can obtain high performance and availability of metadata services in a grid environment.