基于资源聚集的计算网格备份资源选择算法

A Backup Resource Selection Algorithm Based on Resource Clustering in Computational Grid

资源备份是提高计算网格应用可用性的重要方法 .如何为应用选择备份资源是网格资源备份服务要解决的首要问题 .文章提出了基于资源聚集的备份资源选择算法 .该算法将为应用分配的资源按照资源之间的关系聚集成多个资源集合 ,然后根据应用的可用性需求为每个资源集合选择备份资源 ,每个资源集合中的资源共享同一组备份资源 .这一算法适用于计算网格 ,可以在资源备份服务模块中实现 .最后 ,给出了该算法的应用实例 .

Resource backup is an important method for improving the availability of applications in computational grid. How to choose the backup resources is a challenging issue for grid middleware to provide resource backup service. In this paper, we propose a backup resource selection algorithm base on resource clustering. It divides the resources allocated for the application into some resource sets according to the relationship between them, and chooses backup resources for each set based on the availability requirement of the application. The resources in a set can share the backup resources chosen for the whole set. This algorithm reflects the feature of resource allocation and is easy to implement in the resource management architecture of computational grid. At the end of this paper, we show how to use this algorithm to choose backup resources for a real application.