Abstract IoT video applications such as public transportation, traffic monitoring, parking and surveillance systems in smart cities collect and transmit video over multi-hop wireless network fabrics which are known to be less reliable and responsive as their wired counterparts. While multipath routing has emerged as an effective strategy to help meet IoT video transmission’s, one of the main challenges current multipath routing approaches face is to be able to select routes that meet the requirements of the driving video applications in a timely manner. To fill this gap, we recently proposed FITPATH, an adaptive multipath selection mechanism for IoT video transmission that uses a heuristic-based iterative optimization approach to estimate in real time the conditions of the underlying network while accounting for the different bitrate requirements of the application video flows. To provide a complete understanding of FITPATH’s performance, limitations, and applicability in realistic IoT video transmission scenarios, this work presents a comprehensive experimental evaluation of FITPATH under a wide range of conditions. We demonstrate that FITPATH consistently outperforms existing multipath selection mechanisms in terms of both end-user Quality of Experience (QoE) and overall network performance. We also evaluate the performance and convergence of FITPATH’s heuristic, showing that it quickly generates feasible solutions and incrementally improves QoE as the algorithm iterates. Furthermore, we propose practical deployment strategies for FITPATH under both centralized and decentralized network control plane architectures. Additionally, we analyze the impact of control messages and network dynamics on video quality in decentralized network control plane environments.
Bhering et al. (Mon,) studied this question.