Improving Data Access Performance in Cluster Based Ad-Hoc Networks

Abstract

The objective of cooperative caching is to improve data availability, improve access efficiency and reduce query delay in mobile Ad-Hoc networks. Many types of cache replacement algorithms like LRU, LFU, LFRU, LRU-MIN and LFU-MIN are used to improve data accessibility and reduce query delay in cluster based cooperative caching in Mobile Ad-Hoc networks. But they have some limitations such as accessing remote information station via multi hop communication leads to longer query latency and causes high energy consumption, many clients frequently access the database server they cause a high load on the server and reduce the server response time .Multi hop communication causes the network capacity degrades when network partition occurs. The paper gives an overview of Cooperative Cache Management Techniques and caching policies and propose a new algorithm can be regarded as a LRFU-MIN (least recently frequently used information with minimal number of page replacements). It discover a data source which induces less communication cost of moving cache blocks into the most recently frequently used position and minimizes caching duplications between neighbor nodes. In this paper we utilize a cross-layer design approach to improve the performance of combined cooperative caching and prefetching schemes. The paper examines the performance using NS-2 simulation environments. The proposed LRFU-MIN enhances the performance of cross-layer cluster based cooperative caching in mobile Ad-Hoc networks when compared with LRU and LFU-MIN.