A novel QoS negotiation model based on intelligent learning technique in clouds

As more and more high-end applications have been deployed on cloud platforms, the quality-of-service (QoS) of cloud platform has become an important issue that need to be addressed. To achieve desirable QoS, an efficient negotiation model plays a critical role because cloud application often requires plenty of resource during their execution. In this paper, we propose a novel QoS negotiation model which applies intelligent learning technique to adjust the negotiation policy, in this way, QoS agreement between application and resource provider can be quickly achieved even when multiple resource requests are involved concurrently. The proposed negotiation model is implemented in an integrated QoS negotiation framework, in which negotiation agents help to make resource reservations while allocations agents are responsible for committing those reservation requests. The proposed framework is tested in a real cloud platform, and the results indicate that it can significantly improve the negotiation efficiency comparing with other negotiation methods. In addition, the experimental results also show that by using the proposed negotiation model, large-scale workflow applications can reduce their makespan by about 7%~11%.