An empirical model of adaptive cloud resource provisioning with speculation

Cloud computing is a utility model that offers everything as a service and supports dynamical resource provisioning and auto-scaling in data center. The proper load balancing and dynamic resource provisioning improves cloud performance and attracts the cloud users. This impacts a need for adaptive and automated provisioning of resources, aligned with clients’ Service Level Agreement (SLA) amidst the time variant cloud environment. The focus of our work is to study how speculative analysis can be used to predict exact resources for an application, whose accuracy demands solution for under/over-utilization of the resource. We have tested our simulator with varying resource load and proved that our system reduces resource allocation latencies and SLA violations. Experimental results show that our proposed model offers more adaptive resource provisioning, as compared to heuristic and other machine learning algorithms. Our experimental results demonstrate adaptive resource allocation over customer-driven service management, under the rapidly changing requirements of cloud computing.