A Multi-objective Virtual Machine Scheduling Algorithm in Fault Tolerance Aware Cloud Environments

In modern cloud datacenters, virtual machine (VM) scheduling is a complex problem, especially taking consideration of the factor of service reliability. Failures may occur on physical servers while they are running cloud users’ applications. To provide high-reliability service, cloud providers can adopt some fault tolerance techniques, which will influence performance criteria of VM scheduling, such as the actual execution time and users’ expenditure. However, only few studies consider fault tolerance and its influence. In this paper, we investigate fault tolerance aware VM scheduling problem and formulate it as a bi-objective optimization model with quality of service (QoS) constraints. The proposed model tries to minimize users’ total expenditure and, at the same time maximize the successful execution rate of their VM requests. The both objectives are important concerns for users to improve their satisfactions, which can offer them sufficient incentives to stay and play in the clouds and keep the cloud ecosystem sustainable. Based on a defined cost efficiency factor, a heuristic algorithm is then developed. Experimental results show that, indeed, fault tolerance significantly influences some performance criteria of VM scheduling and the developed algorithm can decrease users’ expenditure, improve successful execution rate of their VM requests and thus perform better under fault tolerance aware cloud environments.