Towards SDN-based fog computing: MQTT broker virtualization for effective and reliable delivery

Performance of data analytics in Internet of Things (IoTs) depends on effective transport services offered by the underlying network. Fog computing enables independent data-plane computational features at the edge-switches, which serves as a platform for performing certain critical analytics required at the IoT source. To this end, in this paper, we implement a working prototype of Fog computing node based on Software-Defined Networking (SDN). Message Queuing Telemetry Transport (MQTT) is chosen as the candidate IoT protocol that transports data generated from IoT devices (a:k:a: MQTT publishers) to a remote host (called MQTT broker). We implement the MQTT broker functionalities integrated at the edge-switches, that serves as a platform to perform simple message-based analytics at the switches, and also deliver messages in a reliable manner to the end-host for post-delivery analytics. We mathematically validate the improved delivery performance as offered by the proposed switch-embedded brokers.