Mobility- Aware Cache Management in Wireless Environment

Abstract

In infrastructure wireless environments, a base station provides communication links between mobile client and remote servers. Placing a proxy cache at the base station is an effective way of managing the wireless Internet bandwidth efficiently. However, in the situation of non‐uniform heavy traffic, requests of all the mobile clients in the service area of the base station may cause overload in the cache. If the proxy cache has to release some cache space for the new mobile client in the environment, overload occurs. In this paper, we propose a novel cache management strategy to decrease the penalty of overloaded traffic on the proxy and to reduce the number of remote accesses by increasing the cache hit ratio. We predict the number of overload ahead of time based on its history and adapt the cache for the heavy traffic to be able to provide continuous and fair service to the current mobile clients and incoming ones. We have tested the algorithms over a real implementation of the cache management system in presence of fault tolerance and security. In our cache replacement algorithm, mobility of the clients, predicted overload number, size of the cached packets and their access frequencies are considered altogether. Performance results show that our cache management strategy outperforms the existing policies with less number of overloads and higher cache hit ratio.