End-to-End Response Time Analysis for RT-MQTT: Trajectory Approach versus Holistic Approach

Nowadays, custom components are increasingly being replaced by commercially available off-the-shelf hardware and standard protocols. Additionally, emerging industrial paradigms like Industry 4.0 and IoT place new demands on requirements like scalability, transparency, adaptability and efficiency. Accordingly, application layer protocols like the Message Queuing Telemetry Transport Protocol (MQTT) are becoming more and more popular in these fields, thanks to their simplicity, scalability, low resource-usage and decoupling between end nodes. However, these protocols are not deterministic, thus being unsuitable for real-time applications. Recently the authors proposed a set of extensions to the MQTT protocol, allowing applications to explicitly specify real-time requirements that are then used by a resource manager, implemented in Software Defined Networking (SDN), to create real-time channels. This paper extends the work, providing worst-case analysis using the Holistic and Trajectory approaches. The paper also includes a set of experimental results aiming to verify the correctness of both analysis and evaluate its performance in several scenarios.