Provisioning latency-guaranteed real-time service in wired and wireless networks

Abstract

With rapid development of communication technologies, the demand for real-time applications has grown extensively. Accordingly, such applications require more stringent performance requirements, particularly with respect to latency. In fact, latency performance is a critical factor in enhancing quality of user experience (QoE) provided by the real-time applications, making its improvement and guarantee a high priority for related communication technologies. Improving latency performance, however, is challenging due to the influence of diverse factors, such as environmental conditions and traffic characteristics. In particular, the characteristics of the latency performance in wired networks differ significantly from those in wireless networks. Therefore, it is essential to consider such various characteristics for improving latency performance tailored to each environment and ensuring seamless operation of real-time services. In such a vein, this dissertation introduces approaches for provisioning latency-guaranteed real- time services in wired and wireless networks. First, the dissertation proposes the general-topology cut-through switching (GCS) network architecture and an end-to-end latency prediction model in the GCS network to ensure the latency performance in wired networks. Then, the dissertation introduces the WLAN-assisted AR admission and scheduling framework for latency performance guarantee in Wi-Fi based wireless networks. Lastly, the dissertation experimentally investigates Wi-Fi 7’s non-simultaneous transmit and receive (NSTR) multi-link operation (MLO) to enhance the latency performance of future Wi-Fi devices. Through extensive analyses and experiments, the dissertation demonstrates that the proposed approaches can significantly enhance and guarantee the latency performance of real-time services across wired and wireless networks.