Indexed on: 28 Dec '15Published on: 28 Dec '15Published in: Computer Science - Distributed; Parallel; and Cluster Computing
Cloud computing services are becoming more and more popular. However, the high concentration of data and services on the clouds make them attractive targets for various security attacks, including DoS, data theft, and privacy attacks. Additionally, cloud providers may fail to comply with service level agreement in terms of performance, availability, and security guarantees. Moreover, users may choose to utilize public cloud services from multiple vendors for various reasons including fault tolerance and availability. Therefore, it is of paramount importance to have secure and efficient mechanisms that enable users to transparently copy and move their data from one provider to another. In this paper, we explore the state of the art inter cloud migration techniques and identify the potential security threats in the scope of Hadoop Distributed File System HDFS. We propose an inter cloud data migration mechanism that offers better security guarantees and faster response time for migrating large scale data files in cloud database management systems. The proposed approach enhances the data security processes used to achieve secure data migration between cloud nodes thus improves applications response time and throughput. The performance of the proposed approach is validated by measuring its impact on response time and throughput, and comparing the performance to that of other techniques in the literature.