Performance Comparison of Majority Voting with ROWA Replication Method over PlanetLab.

Since the Web started in 1990, it has shown an exponential growth. It is essential that the Web's scalability and performance keep up with increased demand and expectations. The key to achieving these goals of scalability, robustness and responsiveness lies in the practices of caching and replication. Quorum Consensus is a popular protocol used for data replication. This paper describes an implementation of two special cases of Quorum Consensus protocol, namely Majority Voting and Read-One-Write-All (ROWA) and compares their performance. The performance evaluation was done using a number of systems located at PlanetLab member institutions at different locations over the world. This enabled simulation of real world Internet conditions. The study shows that the ROWA protocol performs better than the Majority Voting under no-site-failure conditions in terms of response time, communication overhead and growing number of users. [ABSTRACT FROM AUTHOR]