WARM Based Data Pre-fetching and Cache Replacement Strategies for Location Dependent Information System in Wireless Environment

Abstract

Rapid growth in wireless communications gives mobile users the facility to search local emergency services, business events, entertainment activity and other necessary information at anytime from anywhere using any device. Recent research on mobility and spatial---temporal data focuses on techniques for improving data availability and reducing latency and access costs through client side data caching techniques. In this paper, we design a cache management system for location dependent information system by considering various contextual information of users such as mobility pattern and the type of service being requested. Cache placement and invalidation is carried out considering the spatial and temporal validity of the services. We also propose a pre-fetching technique which considers the geographical and semantic adjacency of the queried items. The pre-fetching policy rules are derived from applying Weighted Association Rule Mining to determine the secondary service item that should be pre-fetched while processing the primary service query. A new cache replacement algorithm Spatial and Temporal Valid Scope which takes into account the spatial and temporal valid scopes, data distance and service type are applied for eviction. The experimental evaluations using synthetic datasets show that the proposed cache management system is effective in improving the system performance in terms of the cache hit ratio of mobile clients.