Business-driv en Adaptive QoS Optimization Mechanism for Virtualized Multi-tire IT Service

In view of the adaptive QoS Optimization issue for virtualized multi-tire IT Service, basing IT resource regulation only on QoS parameters leads to suboptimal decision. This paper proposes an adaptive online QoS optimization mechanism for multi-tier IT service system from business perspective, and takes the global business utility as the objective of QoS optimization. Compared to traditional QoS optimization approaches, a novel business utility analytical model is built with customer behavior analysis, which more precisely captures the net business revenue. The cost due to IT resource virtualization is also considered in multi-tier virtualized IT service environment. Then the detail optimization algorithm is proposed based on the novel IT-business metrics-mapping model. This approach can adaptively maintain the QoS in a reasonable range by dynamically optimizing the distribution of virtualized IT resource driven by business utility. The experimental simulation shows that this approach is superior to the conventional approach since the overall business utility of all service levels stays on a relative high level.