基于效用的云计算容错策略和模型
摘要
Cloud computing has emerged as a widely accepted computing paradigm over the past few years. With the increasing application of cloud computing in vast areas such as industry, transportation and healthcare, the reliability of application systems draws special attention. However, due to the complexity of cloud computing system in structural and behavioral aspects, it is a great challenge to guarantee the reliability of applications in cloud environment. In this paper, we investigate the fault tolerant strategy and model for cloud computing. By constructing the extensible fault models of cloud computing systems, we can characterize the mechanisms of cloud computing, the fault behaviors of component, the cooperation and competition relations between cloud applications. Considering the fault and service resource of cloud computing, we solve the Nash equilibrium of the model and optimize the recovery program of cloud computing and propose the fault migration and restoration schemas for cloud computing. By considering the fault time and its cost, we compute the utility of the components of cloud computing and cloud application, and further analyze the benefits of cloud application. In order to optimize the fault tolerant effectiveness of cloud computing, we solve the Nash equilibrium of model. Model checking techniques are used to verify the correctness of fault diagnosis and recovery. This research is of both theoretical and practical significance in revealing structural and behavioral characteristics of cloud computing, and improving reliability of cloud computing systems.