Abstract
Empowered by computing at the network edge, data sensed from Internet of Things (IoT) devices can be processed by various IoT applications stored in the nearby cloudlets to reduce the response time between the IoT devices. Together with the recently wide- studied network slicing and Network Function Virtualization (NFV) paradigm, IoT edge computing function can satisfy the individual requirements on the flexibility, security as well as accessibility of the IoT applications. Although the significant research effort made on NFV architectures has received remarkable achievements recently, seldom work has put its emphasis on optimizing IoT edge computing when considering it as a service chain. Since edge computing is regarded to play as a critical role in the context of IoT paradigm to enhance network efficiency, we consider in this paper the dynamic VNF service chain allocating optimization for IoT edge computing. To this end, we detail an Integer Linear Programming (ILP) framework designed to VNF based IoT edge-chains and propose a heuristic scheme to reduce the computation difficulty for large size problems. At last, the attainable system performance of the proposed scheme has been validated by a wide set of investigations.