Data Replication is a key factor for improving the availability of data in distributed database systems. Concurrency Control of replicated data is very important because replicated copies must behave like a single copy ; i.e. mutual and internal consistency must be preserved. Many concurrency control algorithms for replicated data have been proposed. Two of the major families of such algorithms are the Voting Approach and the Special Copy Approach. In this thesis, we examine two algorithms by simulation : Majority Consensus algorithm for the Voting Approach and Available Copy algorithm for the Special Copy Approach. For the simulation, we as sume full replication of database at all sites and no site or network failures. The simulation result shows three facts as follows. First, Available Copy algorithm has better success rate than Majority Consensus algorithm in update transaction. Second, the Performance of Majority Consensus algorithm is much influenced by network capacity. Third, the performance of Available Copy algorithm is much influenced by disk access rate and the number of disks in a site.
