Optimal Dynamic Reserved Bandwidth Allocation for Cloud-Integrated Cyber-Physical Systems

Cloud-integrated cyber-physical system (CCPS) is playing an increasingly important role in our daily life. Unfortunately, how to dynamically make the reserved bandwidth allocation for CCPS is still a great challenge. To address this issue, we propose a novel strategy that systematically and dynamically allocates reserved bandwidth for multiple CPS services in CCPS with a nearly optimal method. By making a balance between the three key factors of crash cost, QoS loss and operating cost, the proposed strategy is able to minimize the operating cost, the influence of the discrepancy between the service-level agreement and the bandwidth actually delivered, while preserving essential QoS level. Then we develop an online algorithm based on the proposed strategy using the Lyapunov optimization theory. The online algorithm can approximate the optimal solution within provable bounds and is capable of processing the tasks within a preset delay. Theoretical analysis of performance proves advantages and shows that the algorithm has the ability to solve the complicated dynamic bandwidth allocation problems for multiple CPS services in practical CCPS. Extensive experiments validate its effectiveness as well as its superiority to five existing strategies (FM, MS + LB, MS + EF, DS + LB, DS + EF) in overall cost, crash cost, QoS loss, and operating cost.