An Adaptive Management Approach Collaborating Heterogeneous Resources in Cloud Service Environments

Abstract

As cloud computing grows rapidly and virtualization techniques become widely-used, it is critical and important to allocate limited resources to various applications on demand for the cloud service environments. In this paper, we propose an adaptive resource management approach considering collaboratively allocating heterogeneous resources to fully utilize extra resource capacity. The system architecture is illustrated, and then the definition of the optimization problem concerning collaborative resource allocation is presented. An optimization algorithm is developed and described, which carries out stochastic and directional search step by step to jointly schedule different resources. The evaluation results of simulation experiments demonstrate that by using the collaborative allocation approach we designed, the performance of different applications deployed in the cloud environment could be guaranteed subject to the QoS (Quality of Service) specification, despite of the significant fluctuation of workloads.