Edge Network Slicing With Statistical QoS Provisioning

Edge network slicing has been identified as a crucial technology to provide dedicated virtual networks with a series of tailored resources (i.e., communication, computing, caching, and backhaul resource) toward differential services. A few critical challenges also arise from the network slicing, especially taking the unknown time-varying environment information and the interdependent communication, computation and caching in account. This letter proposes a novel end-to-end effective capacity (EC) for the edge network, to analyze the long-term performance of edge network slicing. Based on the EC, a resource orchestration mechanism is developed to optimally allocate these heterogeneous resources for the minimization of the network cost under the quality-of-service (QoS) guarantee. Simulation results reveal the synergy among heterogeneous resources in terms of the end-to-end capacity, and also demonstrate the effectiveness of the proposed scheme compared with other schemes.