A Software-Defined Solution for Managing Fog Computing Resources in Sensor Networks
share
The fast growth of Internet-connected embedded devices demands for new capabilities at the network edge. These new capabilities are local processing, fast communications, and resource virtualization. The current work aims to address the previous capabilities by designing and deploying a new proposal, which offers on-demand activation of offline IoT fog computing assets via a Software Defined Networking (SDN) based solution combined with containerization and sensor virtualization. We present and discuss performance and functional outcomes from emulated tests made on our proposal. Analysing the performance results, the system latency has two parts. The first part is about the delay induced by limitations on the networking resources. The second part of the system latency is due to the on-demand activation of the required processing resources, which are initially powered off towards a more sustainable system operation. In addition, analysing the functional results, when a real IoT protocol is used, we evidence our proposal viability to be deployed with the necessary orchestration in distributed scenarios involving embedded devices, actuators, controllers, and brokers at the network edge.
READ FULL TEXT