Grid services performance tuning in OGSA

Abstract

In the next generation of Open Grid Service Architecture - OGSA, Grid has become a Service Oriented environment which is more open, more scalable and more flexible. However, the increase of flexibility will inevitably bring performance decreasing. And the performance problem increasingly constitutes a great challenge to the Grid Researching field. In this paper, we propose a time-based model in order to analyze the performance of Grid Services. On the foundation of this model, we give two methods - Grid Service Instance Pool and Caching on Grid Server - both of them can enhance the performance of Grid Services prominently. Moreover, we not only give theoretic explanation of those methods, but also carry out some experiments by using Globus 3.0 alpha 2 for verification. © Springer-Verlag Berlin Heidelberg 2003.