Binary vote assignment on grid quorum replication technique with association rule

One of the biggest challenges that data grids users have to face today relates to the improvement of the data management. Organizations need to provide current data to users who may be geographically remote and to handle a volume of requests of data distributed around multiple sites in distributed environment. Therefore, the storage, availability, and consistency are important issues to be addressed to allow efficient and safe data access from many different sites. One way to effectively cope with these challenges is to rely on the replication technique. Replication is a useful technique for distributed database systems. Through this technique, a data can be accessed from multiple locations. Thus, replication increases data availability and accessibility to users. When one site fails, user still can access the same data at another site. Techniques such as Read-One-Write-All (ROWA), Hierarchical Replication Scheme (HRS) and Branch Replication Scheme (BRS) are the popular techniques being used for replication and data management. However, these techniques have its weaknesses in terms of communication costs that is the total replication servers needed to replicate the data. Furthermore, these techniques also do not consider the correlation between data during the fragmentation process. The knowledge about data correlation can be extracted from historical data using techniques of the data mining field. Without proper strategies, replication increases job execution time. In this research, the some-data-to-some-sites scheme called Binary Vote Assignment on Grid Quorum with Association (BV AGQAR) is proposed to manage replication for meaningful fragmented data in distributed database environment with low communication cost and processing time for a transaction. The main feature of BV AGQ-AR is that the technique integrates replication and data mining technique allowing meaningful extraction of knowledge from large data sets. Performance of the BVAGQ-AR technique comprised the following steps. First step is mining the data by using Apriori algorithm from Association Rules. It is used to discover the correlation between data. For the second step, the database is fragmented based on the data mining analysis results. This technique is executed to make sure data replication can be effectively done while saving cost. Then, the databases that are resulted after the fragmentation process are allocated at their assigned sites. Finally, after allocation process, each site has a database file and ready for any transaction and replication process. Finally, the result of the experiments shows that BV AGQ-AR can preserve the data consistency with the lowest communication cost and processing time for a transaction as compared to BCSA, PRA, ROW A, HRS and BRS