Enhancing MQTT with Real-Time and Reliable Communication Services

Abstract

MQTT is an application-layer protocol that eventually became popular in the Internet of Things (IoT) and Industrial IoT (IIoT) thanks to its simplicity and effective publisher-subscriber messaging model that enables its use in embedded resource-constrained devices. However, MQTT features a limited set of Quality-of-Service classes addressing exclusively message delivery, impairing its use in IIoT applications subject to timeliness requirements. This limitation of MQTT has been addressed in the literature, but with focus on the broker real-time operation, only. This paper adds to the state-of-the-art, by proposing a set of extensions to the MQTT protocol grounded on Software-Defined Networking (SDN) that enable, at the network level, attaining real-time communication services. Simulation results validate the benefits of the proposed extensions.

A Scalable Real-Time SDN-Based MQTT Framework for Industrial Applications

“…The aim of RT-MQTT [12], [13] is to provide real-time guarantees to time-sensitive (TS) MQTT flows, while keeping full compliance with the standard. To this end, the realtime requirements associated with a particular TS topic are conveyed within the User Properties of MQTT messages.…”

“…Although real-time computing systems have found extensive use in industrial domains [59], [60], analyzing the worstcase timing behavior of communication protocols can be a challenging endeavor. This section firstly refines the prior analytical approach for TS flows within the RT-MQTT architecture [15], with a focus on enhancing system analyzability and predictability. Then it introduces the adaptation of HA and TA to the MRT-MQTT architecture, enabling to derive end-to-end timeliness guarantees for TS flows in edge networks.…”

“…The final term in the equation is computed considering the start time of message m i transmission at the output port last i . Consequently, (12) provides the worst-case response time for m i given by TA.…”

“…end if end for end for end while last switch visited and subsequently calculates the worst-case response time for m i (ResponseT imeCalc(i) in accordance with (12)). The termination condition for the algorithm is met when there exists a message for which the end-to-end response time surpasses its predefined end-to-end deadline.…”

“…• Developing an SDN-based resource management framework along with proposing a set of real-time extensions to MQTT, referred to as RT-MQTT [12], [13]. This architecture empowers applications to specify realtime MQTT requirements, which are then translated into SDN/OpenFlow-enforced network reservations to establish real-time channels; • Evaluating RT-MQTT system predictability and schedulability through response time analysis [14], [15]; • Enhancing the scalability of RT-MQTT with a new architecture, known as MRT-MQTT [16], that leverages multicast-based connectivity across multiple edge networks that follow the RT-MQTT architecture. This paper extends this body of research in several significant ways:…”