A mobile agent-based statistic execution model for grid computing

Abstract

This paper employs mobile agents in grid computing systems. We propose a hierarchical routing model by using mobile agents to collect and update the dynamic changes of routing information in grid systems. User tasks are sent from the source node to the destination in a statistic manner, which has been proved to be the only optimal solution that makes inference on the known traffic information and approximates to a unbiased distribution. Our method is fault tolerant, scalable, and relies completely on local information. Extensive simulations shows the prior of our algorithm to state-of-art algorithms. © 2008 Springer-Verlag Berlin Heidelberg.