Optimizing the performance of Database as a Service (DaaS) model — A distributed approach

Database as a Service (DaaS) is one among the latest service models offered by cloud computing and is adapted by most of the organizations. Database fragmentation, replication and allocation are the key issues to be tackled by most of the DaaS framework. Out of these issues the problem of replication is discussed throughout this paper. Optimizing the performance of DaaS service model by means of strategic database replication is the motivation behind this paper. We assume that the database fragmentation is already done for a given DaaS platform, and we take database fragments as the unit of replication in our replication procedure. As part of the survey, we studied the existing database replication schemes in detail and identified their major performance issues. After review we introduced a new strategy for database replication in DaaS framework. Our approach solves some of the current issues involved and targets to mixed goals out of replication. The replication strategy proceed in three phases; selecting the most popular database fragment, its placement at optimal server and finally the maintenance of replicas forever. Simulation experiments are conducted to evaluate the performance of proposed algorithm. We observed that our algorithm runs in polynomial time and yields suboptimal solution.