Apply IBM ObjectGrid Technology To Extreme High Volume Data Processing Systems

Extreme High Volume Data Processing Systems are systems that process high volume data with restrict requirements of performance, scalability, availability, and security, which are usually implemented by Distributed Computing. At present, Distributed Computing technology is classified to web service, middleware and grid computing. Data Grid has gotten a biggish development as one of the research areas in grid computing which mainly focuses on the storage and computation of large scale distributed data. It supplies the storage, management, transmission, access and local operation of the distributed data. IBM ObjectGrid is one of the newest implementation of data grid technology to support java object in data grid which provides solutions for Extreme High Volume Data Processing Systems.This paper introduces the architecture and programming model of IBM ObjectGrid; summarizes the application models to distributed memory, high available singleton and partition of data and computation; applies each model to suitable applications; explains the technical problems and the solutions in each application model. At last, this paper uses the equity cross system to give a demonstration how to apply the application models of IBM ObjectGrid to real High Volume Data Processing Systems. It describes the combination of distributed memory model and IBM WPF technology at first, then replaces the IBM WPF with high available singleton, uses partition of data and computation to realize the data storage and computation base on partition. IBM ObjectGrid increases the performance of the equity cross system for some applications that will use ObjectGrid, there may be some other issues to be considered, including the security of data in the ObjectGrid and performance tuning. These issues will be researched in the future work.