The Publish/Subscribe communication pattern has proved to be particularly tailored to the IoT world, with the MQTT protocol being the nowadays standard de-facto for IoT applications. Request/response protocols explicitly designed for the IoT, such as CoAP, have been revised to support also Publish/Subscribe. The purpose of this paper is to perform a comparison between two protocols: MQTT-SN, the version of MQTT thought specifically for sensor networks, and CoAP in its Pub/Sub version, defined in a recent IETF draft. Both protocols are Pub/Sub in nature and based on UDP at the transport layer, allowing therefore a fair comparison of their functionalities. We propose a open-source implementation of the CoAP Pub/Sub version and we compare the two protocols: first from a theoretical perspective and, then, in a simulated environment characterized by varying number of clients and network conditions. Results show that CoAP represents a valid alternative to MQTT-SN for publish-subscribe environments; in particular, CoAP results being the best choice for highly dynamic networks.

Internet of Things and wireless sensor network applications are becoming more and more popular these days, supported by new communication technologies and protocols tailored to their specific requirements. This paper focuses on improving the performance of a Wireless Sensor Network operated by the MQTT-SN protocol, one of the most popular publish/subscribe protocols for IoT applications. In particular, we propose a dynamic Quality of Service (QoS) controller for the MQTT-SN protocol, capable of evaluating the status of the underlying network in terms of end-to-end delay and packet error rate, reacting consequently by assigning the best QoS value to a node. We design and implement the QoS controller in a simulated environment based on the ns-3 network emulator, and we perform extensive experiments to prove its effectiveness compared to a non-controlled scenario. The reported results show that, by controlling the quality of service, it is possible to effectively manage the number of packets successfully received by each device and their average latency, to improve the quality of the communication of each end node.

No abstract

No abstract