Energy Efficient Varying Fanout Indexing Technique for Skewed Access Patterns in the Wireless Mobile Environments

Abstract

As we know, due to limited battery power, the most important issue in mobile computing is energy saving, which can be achieved through indexed data organization to broadcast data over wireless channels to a large no of mobile clients. In this paper, we explore the balanced and imbalanced index tree with varying fanout over skewed data. We purpose a varying fanout indexing technique with replication for data broadcast with skewed access pattern over a single wireless communication channel. We also show that replication can be performed at any level in varying fanout index tree, which increases the length of the overall broadcast cycle but reduces the directory miss. We compared our technique with the conventional as well as existing techniques. The performance results suggests the superiority of this technique over another replicated index technique i.e., fixed fanout index in all aspects. Our index technique also ensures correctness of results when larger size of broadcast file is used. From the performance analysis, the proposed indexing technique outperforms fixed fanout index technique.