Computation Offloading Optimization Based on Probabilistic SFC for Mobile Online Gaming in Heterogeneous Network

Abstract

To support new services targeted by 5G, great efforts have been taken not only on the research work of new waveform design and air interface but also on cloudification and softwarization for future heterogeneous network. As one of the most popular services toward 5G, cloud gaming offloads computation-intensive tasks to the cloud in order to alleviate the computation burden of mobile devices, but it introduces latency which deteriorates user experience especially for the delay-sensitive online game. In order to solve the optimization problem of resource allocation with the quality of experience guarantees and reduce the operational expenditures and capital expenditures of mobile operators for deploying online game, fog computing and network function virtualization are deemed as promising solutions. In this paper, a component-based approach is proposed to model online game based on the probabilistic service function chain. In order to obtain the optimal virtual function placement in the fog-enabled heterogeneous radio access network, the cost minimization of computation offloading on the data plane is formulated as an integer linear programming problem considering the constraints of application maximum tolerable latency, resource limitation, and user behavior. The optimization problem is NP-hard. To solve the problem with low complexity, the heuristic algorithm is proposed called Probabilistic Service function chain Embedding based on Cost Optimization(PSECO). The performances of the two algorithms are evaluated. The simulation results show that the costs are affected mainly by the number of components, the arrival rate of user requests, mobile user behavior, as well as the physical network topology and the number of users. The heuristic algorithm PSECO has optimal results with low complexity and it is suitable for large scale networks.