Department of Information Technology (DIT)
Due to missing deadlines, the jobs are being rejected. The rejection rate increases. The authors want to propose on-demand provisioning in order to maintain QoS they need to give extra resource to the job. Mapping performance requirements to the underlying resources in the cloud is challenging. Resource under-provisioning will inevitably hurt performance and resource over-provisioning can result in idle instances, thereby incurring unnecessary costs. There is an increasing demand to efficient management of large-scale application under cost and deadline constrained.