A Self-adaptive QoS-management Framework for Highly Dynamic IoT Networks

Abstract

IoT infrastructure makes great demands on network control methods for dynamic and efficient management of massive amounts of nodes. Software-Defined Networking (SDN) enables to handle dynamically network traffic as well as flexible traffic control in real-time. However, while providing flexibility and scalability, SDN-based architecture still remains ineffective to self-adapt with respect to network topologies with more or less switches in the data plane (highly dynamic topology). Having a centralized control plane is not an acceptable situation because that would represent a single point of failure in the network. Using multiple controllers that ensure flexibility and high availability would be a solution; meaning that if one controller has problems and fails, the other would be ready to take over and control the network. Thus, having a single controller raises the problem of scalability while multiple controllers call for a distributed states management problem. To overcome such issues, we propose EFQM++, a selfadaptive framework for highly dynamic network topology changes. By leveraging SDN controller topology discovery mechanism, EFQM++ improves flow end-to-end transmission delay. It tackles flexibility and scalability related to a single point of failure problem and gives distributed states management solutions in large scale IoT networks. EFQM++ reduces up to 6% and 13% the average delay in contrast to previous works like EFQM and AQRA, respectively.