A Session-Based Adaptive Admission Control Approach for Virtualized Application Servers

Abstract

This paper presents a session-based adaptive admission control approach for virtualized application servers called ACVAS (adaptive Admission Control for Virtualized Application Servers). ACVAS uses measured and predicted resource utilizations of a server to make admission control decisions for new user sessions. Instead of using the traditional on-off control, it implements per session admission control, which reduces the risk of over-admission. Moreover, instead of relying only on rejection of new sessions, ACVAS takes benefit of the cloud elasticity, which allows dynamic provisioning of cloud resources. It also implements a simple session deferment mechanism that reduces the number of rejected sessions while increasing session throughput. Thus, each admission control decision has three possible outcomes: admit, defer, or reject. Performance under varying user load is guaranteed by automatic adjustment and tuning of the admission control mechanism. The proposed approach is demonstrated in a discrete-event simulation.