Study On Grid Data Management Based On DataSpaces

The fast development of Grid technology especially Data Grid takes lots of data recently.These data aim to share with others under the characters of heterogeneous and distribute, which can't be solved suitably by traditional data management theory. This results in a jungle of data processing solutions and a jungle of data and file formats.How to manage and use these heterogeneous data becoming one of the most important problems of data management research, also of Grid data management research.Dataspace is a new concept of information management for these new chanllenges faced by data management research proposed by M.Franklin in 2005. It contains a number of heterogeneous data sources and partially unstructured data. Dataspace considers all the data in it as participants using an abstract data model, then provides these participants a lot of services including catalog, index, discovering and other managements. Also dataspace has to integrate the data in pay-as-you-go fashion.Some researches have improved that the theory can solve new problems of data management effectively.With combining the technologies of data management and data Grid,we propose the architecture of dataspace management system to manage the Grid data base on the Grid middleware named Globus Toolkit.This paper describes the details of implement of Grid DataSpace Management System(GDSMS) based on OGSA-DAI. We using work-flow and resource-service, which are characteristic of the Grid data access middleware, to design and then to implement the prototype system of GDSMS.The contents of the paper are as follow:Discuss the meaning of integration of Grid data research; introduce the status of research of Grid,Data Grid and also relative middleware.Summarize of dataspace research status. Analyse the technology relative with dataspace, including data model,indexing and integration.Design the GDSMS system with analysing the details of each component.Implement the prototype system of GDSMS using OGSA-DAI based on Globus Tookit 4.0. It is the foundation for future work of research.