Performance Evaluation of Video Streaming Application Over CoAP in IoT

Abstract

The Internet of Things (IoT) is considered a milestone for the realization of intelligent services through the interaction of smart things. Constrained application protocol (CoAP) was proposed by the Internet engineering task force as a web transfer protocol in the IoT. CoAP inherits the properties of the representational state transfer model to realize communications with low overhead. It is characterized by a small overhead and it uses the stateless UDP protocol in the transport layer. As the IoT has attracted more attention, the demands for various multimedia applications based on the IoT have increased. Video applications and IoT objects communicate through CoAP. CoAP defines a simple congestion control mechanism that can provide reliable communication between IoT endpoints. However, the congestion control mechanism was not proposed considering video streaming applications. In this paper, we evaluate the performance of video streaming applications for reliable CoAP communications between devices over a wireless network. To evaluate the performance, we perform multiple experiments by adjusting the CoAP transmission parameters. Moreover, we analyzed the effect of the segment duration of a video being streamed over CoAP. The results show that the default CoAP parameters do not fulfill the stringent Quality of Experience constraints. The experiments show that the default retransmission RTO is extremely high for the video streaming application. A small increase in the packet loss can significantly affect the video download time which increases the risk of playback interruption. Through experiments, we show that the performance of video streaming applications can be improved by tuning the congestion control parameters according to the network conditions. A smaller RTO value can ensure significant reduction in the segment download time. Similarly, the video can be downloaded quickly by increasing the number of unacknowledged transactions. Furthermore, the experiments show that the risk of playback interruption can be mitigated by downloading longer segments.