Queuing Theory-Based Modeling and Optimization of a Publish/Subscribe IoT Communication System

This paper presents a comprehensive Queuing Theory-based model of a Publish/Subscribe Internet of Things (IoT) communication system. Using a Markovian (<tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathrm{M} / \mathrm{M} / 1$</tex>) queuing process, we model, analyze, and optimize the message delivery of a Middleware Message Queuing Protocol (MMQP)-broker. We derive key performance metrics such as average waiting time, system utilization, and the effects of multi-threading on performance. Our model aims to optimize the throughput and the efficiency of the system by determining the optimal service rate of the broker in relation to the overall arrival rate of messages. Through experimental evaluation, we demonstrate the model's accuracy and its relevance in providing valuable insights for improving IoT communication.