Research on Adaptive QoS Weakly Hard Real-time Schedule Algorithm

摘要

Real-time networked applications are widely used today. Quality of service (QoS) assurance is a major concern in real-time networked applications. Especially, how to degrade the QoS gracefully in overload conditions is still a challenging problem. Typically, weakly hard real-time scheduling algorithms and admission controllers are key approaches to guarantee QoS. In this paper, both methods are analyzed. Unfortunately, the existed scheduling algorithms have little adaptability and admission controllers cause under-utilization of the system bandwidth. So we present an adaptive QoS weakly hard real-time scheduling algorithm, which combines DBP policy and KWCS policy together and uses proper policy according to run-time overload state. Results show that the mixed algorithm has the advantage of both DBP and KWCS, and it can adoptively degrade the QoS gracefully.