A resource discovery tree using bitmap for grids

There are possibly lots of resources in grid environments today. These resources may be geographically distributed and hard to locate. Thus, finding appropriate resources, that satisfy users' requests, is an important task. Traditional centralized resource discovery mechanisms suffer from single points of failure and heavy load, and are thus not suitable for large numbers of resources. In this paper, we propose a tree architecture for resource discovery. All resources and user queries are transformed into a bitmap index representation. Each leaf node in the tree will store the information about its local resources. Each Index Server (IS) will store the information about its local resources and the information about its children nodes. The index in each IS is the bitmap of bitwise OR of all its children nodes. We use the AND operation to check whether a node has the resources to satisfy user's requests. We evaluate the performance of our proposed algorithm by comparing it with other algorithms. The experimental results show the number of visited nodes and performed operations, as a result of query forwarding, are smaller than in comparable approaches. Furthermore, the cost of update in our proposed algorithm is also lower. Therefore, our algorithm can effectively improve the system efficiency for resource discovery. (C) 2009 Elsevier B.V. All rights reserved.