Performance Analysis of Cache Consistency Maintenance in Mobile Environment using Agent Technique

Abstract

This paper illustrates three Agents Based Techniques namely, Thread Based Cache Consistency Model, Multiple Server Cache Consistency Model, and Cluster Based Cache Consistency Models used to solve the problems of cache consistency maintenance and also compared with existing method. The drawbacks of existing method are latency, excessive bandwidth wastage, query delay, server overloads as well as network traffic. In the Thread Based Cache Consistency Model, the log is maintained in both server and client, it is easy for server as well as client to make their data object reliable. Agents at server and client make an internal request response operation(s) to access or update data object. It is easy to update the cached data and as well to maintain consistency between client and server. The Multiple Server Cache Consistency Models where the node is designed as a middle server (master client cache) between server and clients, some of the client data cache is placed in a middle server. The middle server is near the client, as result the communication cost and energy consumption is comparatively less. It reduces server’s workload. The Cluster Based Cache Consistency Model designed as a Cluster head node between the server and the client using agent technique, effectively reduces the query latency. These three Agent Based Models are implemented in Ns2 and their performance was analysed and compared with existing method. Agent Based Technique resulted in better performance in relation to update delay, through various numbers of nodes and update rate.