Design and Implementation of a COAP-Based Broker for Heterogeneous M2M Applications

摘要

In the Internet of Thing (IoT) vision, Machine-to-Machine (M2M) applications will be able to communicate with remote physical devices, being continuously updated as the status of things and monitored phenomena change. However, IoT environments have strict constraints concerning limited energy, low bandwidth, and limited computing capabilities. Thus, network congestion and scalability issues arise if many sensing applications with heterogeneous requirements demand state updates for the same resource too frequently. To address this problem, in this work we propose, implement and evaluate an application broker for heterogenous M2M applications. In particular, our application broker acts as the proxy between remote applications and constrained devices. Then, it dynamically regulates the frequency of resource state updates leveraging the measured transmission reliability of update notifications from that resource. A key feature of our proposed approach is the use of a max-min fair allocation procedure to obtain a more efficient utilisation of the limited network resources, without starving some applications. A prototype of our application broker has been developed for COAP, a standard web-based transfer protocol for IoT environments. The proposed approach is extensively evaluated through emulation and compared to a conventional proxy that is unaware of network congestion. Our results show that our solution can provide a significant improvement regarding network utilisation and energy consumption, as well as to better meet applications' requirements.