Dynamic and static controller placement in Software-Defined Satellite Networking

摘要

Software-Defined Satellite Networking (SDSN) [1], similar to Software-Defined Networking (SDN) [2], has emerged as a new paradigm that offers the programmability required to dynamically configure and control a satellite network. As SDSN is large-scale network, recent proposals have advocated deploying multiple controllers that work cooperatively to enhance performance and scalability [3,4]. Nonetheless, this approach introduces a new problem: how many controllers are needed, and where should they be deployed? We call this controller placement problem, which can be divided into dynamic and static types. The former assumes that the network operator has already placed plenty of back-up controllers throughout the network. By changing the on off status of them, it realizes adapting the number and location of active controllers with changing network conditions. On the contrary, the latter decides the controller deployment only for once before building SDSN. Related work only focuses on ground network and several important factors affecting the controller placement are not taken into account comprehensively. In this paper, we propose the framework and mechanism of SDSN, define the Dynamic Controller Placement Problem (DCPP) as well as Static Controller Placement Problem (SCPP), and introduce a heuristic method to solve them. Simulation results show that our solution improves the performance of SDSN.