Improved End-to-end service assurance and mathematical modeling of message queuing telemetry transport protocol based massively deployed fully functional devices in smart cities

QoS is an arguable feature of Message Queuing Telemetry Transport Protocol (MQTT) that is employed under constrained environment for communication over an unreliable network in smart cities. It gives extra control to the client for matching its needs according to the application but with confiscated network performance. For a reliable end-to-end service assurance, guaranteed QoS and controlled scalability, MQTT based IoT must be examined for various network parameters. In this work, an improved TCP based transparent MQTT network is proposed for massively deployed connected devices in smart cities. The mathematical modeling is carried out by analyzing end-to-end quality assurance. Network latency, content delivery rate, number of subscribe/publish requests and services offered by the MQTT broker for various topic IDs are then verified on physical network. These parameters are examined over a transparent gateway-based network for verifying the agreement of devised probability based mathematical model with the actual content delivery, service rate and request hold time at virtual machine based MQTT broker, local network broker and remote server respectively. The Improved MQTT model surpasses Apollo, RabbitMQ and Mosquitto server by evidencing the message queueing delay of 3.5 ms for QoS-0 and 3.6 ms for QoS-1 service IoT.