Adaptive Fault Tolerance for Online Tasks Scheduling in Cloud Computing

Abstract

although cloud computing is a new model of computing with a lot of benefits, it has difficult challenges to be resolved in order to enhance the performance of cloud environment. These challenges are related with the load balancing management, fault tolerance and different security issues. In this paper, the main concerns are the load balancing and fault tolerance in cloud computing. Load balancing is the mechanism of distributing the load among various virtual machines of the cloud to avoid the situation where some of the resources are heavily loaded while other resources are idle. There is also a necessity and a growing demand for fault tolerance to fulfill the availability and reliability because the user's services are processed remotely on cloud resources. This scenario may results in more chances of errors and missing control over cloud virtual machines. The main goal of this research is proposing a hill-climbing scheduling algorithm based on load balancing and adaptive virtual machine reliability. Experimental results clarifies that the proposed algorithm reduces the degree of unbalancing between available VMs, increases the overall performance and distributes the tasks on available VMs depending on its reliability.