Optimal fault tolerant service provisioning for cloud application

Providing fault tolerance services to cloud applications is significant. This paper proposes a novel fault tolerance service provisioning scheme. First, fault tolerant requirements is defined from the reliability of cloud application components, application response time and fault tolerance effects. Then, an optimal fault tolerance service provisioning method is developed based on replication, N-version programming and checkpoint technologies, considering whether cloud infrastructure resources are sufficient. Using optimization theory, we compute an optimal fault tolerant service provisioning solution. The experiment results show that our work could better satisfy the requirements for both fault tolerant service provider and cloud applications.