LAUSR

An accurate resource scheduling system (RSS) for virtual machines based on CPU monitoring and load assessment is presented to solve the shortcoming of resource scheduling in cloud computing systems. A new architecture is designed to improve Credit scheduler, including three core components: CPU load monitoring component (CLMA), CPU load assessment component (CLAA), and the resource adjustment component (RSA). On the basis of the prototype design, we make an evaluation between Credit scheduler and our system with a typical example in Xen platform. The experimental results show that the proposed system could satisfy the personalized resources requirements from users with higher tasks resource utilization and lower system resource utilization when compared with Credit scheduler. RSS has a strong sensitivity to meet the requirements of cloud computing systems, since it can accelerate the executions of applications via dynamic resource scheduling.