MQTT – Overview

A server operates as an MQTT message broker handling specific predefined feeds, known as topics, which IoT devices publish their information to. Upon receipt of this information, the message broker forwards it to all current subscribers for that topic. The taxonomic layout represents that of a folder structure for example ‘/Volkswagen/Touran/1.9-TDi’. Subscribing clients are then able to choose which level of information they wish to receive, subscribing to the ‘/Volkswagen’ topic would imply that a subscriber wishes to receive specific information published only on the Volkswagen feed. Whereas, subscribing to ‘/Volkswagen/Touran/1.9-TDi’ would specifically filter out everything other than Touran 1.9TDi feed information. It is also possible for users to subscribe to multiple topics simultaneously by using the ‘*’ wildcard character, for example ‘/Volkswagen/*’ would automatically include all information for all subtopics of the Volkswagen topic.

MQTT (Message Queuing Telemetry Transport ) is a machine-to-machine “Internet of Things” connectivity protocol. It was designed as an extremely lightweight publish/subscribe messaging transport and is useful in low power, low network bandwidth scenarios. It is easy to deploy and provides a robust messaging system with low overheads both in computing power and network usage.


MQTT is perfectly matched to IoT messaging as it was originally designed with reliability and cost-effective remote monitoring in mind and works very well over unreliable connections or in networks with bandwidth issues or fluctuating latency. The MQTT protocol features a buffering process that allows to be temporarily stored for subscribers and then resubmitted in periods of downtime or when connections between the subscriber and broker are severed. Where the publishing device is disconnected or becomes unavailable the broker is able to forward a predefined or cached message to subscribers so that action can be taken to reinstate the device connection.

The original protocol was developed for IBM in 1999 by Arlen Nipper of Arcom and Dr. Andy Standford-Clark of IBM for the monitoring of equipment used to pipe fossil fuels such as gas and oil across great distances through remote locations and has since expanded into an open protocol. Present day implementations vary from Sensorstream’s localised weather monitoring and security alerting systems through to instant messaging with the likes of Facebook recognising MQTTs potential for super efficient global text messaging across mobile platforms.

In Sensorstream’s SWAN (Sensorstream Wide Area Network) trials MQTT has proven to be a superb protocol for monitoring and reporting in real time from remote locations.