Optimized Dynamic Data Replication for an Efficient Placement of Data in a Cloud Environment

Abstract

In cloud computing, it is important to maintain high data availability and the performance of the system. In order to meet these requirements, the concept of data replication is used. Data replication achieves the goal of effectively replicating the same data to different locations with zero loss of information in the event of a zero-downtime failure. Dynamic data replication strategies in the cloud (providing a runtime location for replicas) need to optimize key performance metrics parameters such as average response time, availability, cost, and load balance. The problem is, as the number of replicas of a data file increases, the data availability and the performance also increase, but at the same time, the cost of creating and maintaining new replicas also increases. This problem needs to find appropriate the replica to be selected and create the replica data based on its availability. Then place the replica by optimizing different parameters such as cost, distance, load balancing, etc. Because of this, the thesis proposes an Optimized Dynamic Data Replication for an Efficient Placement of Data in a Cloud Environment, which is a hybrid of Enhanced artificial bee colony (EABC) and Particle swarm optimization (PSO). These algorithms have certain drawbacks. ABC algorithm is the simple technique for data replication but it has the problem of inefficient resource utilization. PSO has a problem with execution time, and throughput when the environment is heterogeneous. The key aim of this thesis is to minimize the average response time and the replication cost, to reduce waiting time and data transmission time, and balance load. The proposed strategy is compared with another data replication approach in order to performance evaluation. A detailed performance evaluation study has been done to validate the proposed strategy. The proposed algorithm was tested and simulated on Cloudsim. Experimental results compared to different familiar algorithms. Based on the experiment results, the proposed strategy has reduced replication cost by 13.85% and waiting time by 18.37% compared to DCR2S, It has also reduce waiting time by 6% and cost reduction by 8.2 %, outperforming GA. And also reduced waiting time by 5.2% and replication cost reduction by 6.1% compared to ABC, and reduced waiting time by 4.7% and replication cost reduction by 5.7%, outperforming PSO