SLA-DO: A SLA-Based Data Distribution Strategy on Multiple Cloud Storage Systems

摘要

One of the major advantages of the cloud storage system is that it simplifies the time-consuming processes of hardware and software provisioning, deployment and distribution for users. Currently there is a tremendous increase in the scale of data generated as well as being consumed by applications on cloud storage systems, but besides availability metrics, few of cloud storage services provide data privacy guarantees in their Service Level Agreements (SLAs), which can be a major concern for the adoption of cloud storage services. The paper firstly proposes an analysis approach for the data privacy, then introduces privacy into the service quality, and defines a service quality evaluation model including privacy, availability, throughput, transfer time and operating cost. At last, a novel data distribution strategy called SLA-DO is presented. The comparison experiment results confirmed that SLA-DO strategy shows better performance in SLAs compliance, resource utilization, security guarantees and multiple cloud environmental adaptability than the data distribution policy adopted in HDFS and OpenStack.