SPRNT: An Aggressive Approach to Dynamically Schedule Resources with Limited Service Failure in VM-Based Data Centers

Effective virtualized resource provisioning is one of the most challenging tasks of the cloud computing system. Especially when visits to an online service are rapidly increasing, state-of-the-art approaches cannot response to the workload changes in time. In this paper we introduce SPRNT, a framework that accelerates the resource provisioning. By using SPRNT, the cloud system can quickly adapt to the explosively increasing workload and ensure that the SLA is not violated. The experimental results show that SPRNT can deal with the tremendously increasing of the workload and can also reduce the over-provisioned resources efficiently and effectively.