A Three Criteria Data Replication Scheme using Data Mining for Wireless Cellular Network

A centralized collaborative system between nodes and Base Stations (BSs) is developed, and a new data replication scheme based on group and query prediction mobility scheme with Data Mining techniques is proposed. Utilizing the group mobility with the properties of data objects this new scheme (in various sites) minimizes the cost of remote access and improves the response time of the queries. A series of algorithms is developed so that the system can provide the “prefetch” operation for the group of users according to their mobility and their shared data have to wait for them in their next appropriate cell. The data have to follow the user moves. Our new approach contains a prediction scheme for Group and Query Mobility (GQM) based on two phases the Merge Itemset Algorithm (MIA) and the Dynamic Replication Scheme with three Criteria (DRS3C) respectively. After the data reservation operation by site, the DRS3C can guarantee the acceleration of the query execution. DRS3C works in cooperation with MIA and Connection Admission Control (CAC) in order to minimize the network messages, and the response time. It can also be used in the design of any new architecture for distributed query processing. Simulation results are provided. Key-Words: Mobile and wireless computing, Data Mining, Mobile Database, Distributed Databases, association rules