云计算系统可靠性研究综述

As a new computing paradigm, cloud computing has attracts extensive concerns from both academic and industrial fields. Based on resource virtualization technology, cloud computing provides users with services in the forms of infrastructure, platform and software in a "pay-as-you-go" manner. In the meanwhile, since cloud computing provides highly scalable computing resources, more and more enterprises and organizations choose cloud computing platforms to deploy their scientific or commercial applications. However, with the increasing number of cloud users, cloud data centers continuously expand and the architecture becomes increasingly complex, leading to growing runtime failures in cloud computing systems. Therefore, how to ensure the system reliability in cloud computing systems with large scale and complex architecture has become a huge challenge. This paper first summarizes various failures in cloud systems, introduces several methods to evaluate the reliability of cloud computing, and describes some key fault management mechanisms. Since fault management techniques inevitably increase energy consumption of cloud systems, this paper reviews current researches on the trade-off between reliability and energy efficiency in cloud computing. In the end, we propose some major challenges in current research of cloud computing reliability and concludes our paper.