Research And Implementation Of Data Management Technology In AMS E-Science Activities

Science data produced by AMS science activities is increasing by the scale of Peta-Byte. How to manage massive and distributed science data efficiently has become a big challenge in AMS science activities. To address this challenge, Data grid is applied into AMS science activities as platform for the integration of data and storage resource. The management of science data based on data grid is the research focus of this paper.Firstly, this paper provides two data management models: a model based on collection and a model based on market economy. Based on these models, a general architecture for E-Science data management is provided. In this architecture, a systematical research has been focused on the key technology: metadata management,replica management and data transport technology. Finally, by integrating all these key technology, a Monte Carol data management prototype system based on data grid has been constructed in LAN environment. This system integrates all kinds of heterogous resource distributed on network, forms a virtual storage environment, provides uniform web access interface for high level application, and achieves the purpose of sharing science data. At the same time, this platform can also be applied into other data-intensive application fields such as astronomy and bioinformatics and support massive data management in these fields.