Self-adaptive QoS-aware resource allocation and reservation management in virtualised environments

As cloud computing develops and enters the commercial domain, the dynamics and complexity of virtualised environments pose some challenges in managing the resources to ensure the quality of service (QoS) continuously met under fluctuating workloads. In this paper, we focus on resource allocation and advance resource reservation management issues for virtualised environments. The architecture of virtualisation-based resource reservation is first described and then the formulation of an optimisation problem concerning the reservation acceptance decision is presented. To solve this problem, a self-adaptive QoS-aware reservation management approach is proposed, which calculates the reservation acceptance gain to make the final decision. Performance evaluation results demonstrate that by using our approach, the QoS of both types of applications could be guaranteed and thus the total revenue of the resource provider could stably keep rising. Detailed analysis shows that proper choices can be made to deal with resource reservation requests via a QoS-aware manner.