Deterministic Scheduling over Wi-Fi 6 using Target Wake Time: An Experimental Approach

TL;DR

This paper addresses the lack of airtime guarantees in dense Wi‑Fi networks by leveraging Target Wake Time (TWT) in Wi‑Fi 6 to implement deterministic channel access. It develops an optimization framework to synthesize TWT schedules that maximize proportional fairness while honoring minimum throughput and uplink/downlink constraints, and it implements these schedules on a testbed using off‑the‑shelf hardware. The results show that TWT‑based deterministic scheduling can improve both individual TWT client throughput and overall network throughput compared to legacy CSMA/CA, even in heterogeneous client environments. The work highlights practical challenges such as firmware variability and proposes a pseudo‑client and overlap control mechanism to manage scheduling in real deployments, pointing to future work on larger, multi‑AP deployments and standardized firmware support.

Abstract

Wi-Fi networks traditionally use Distributed Coordination Function (DCF) that employs CSMA/CA along with the binary backoff mechanism for channel access. This causes unavoidable contention overheads and does not provide performance guarantees. In this work, we outline some issues that occur with the probabilistic channel access in highly congested scenarios and how those can be mitigated using deterministic scheduling. Towards this, we propose to use Target Wake Time (TWT) - a feature introduced in Wi-Fi 6 as a power-saving mechanism, to improve the performance of Wi-Fi. To gain insights into the workings of the TWT over commercially available off-the-shelf components and to analyze the factors that affect its performance, we carry out various experiments with it over our Wi-Fi 6 testbed. Using these insights and analysis, we formulate and solve an optimization problem to synthesize deterministic schedules and obtain the optimal values of various system parameters. Lastly, we configure our testbed with these optimal parameter values and show that the TWT based deterministic scheduling consistently results in better performance of the TWT-capable clients and overall system performance compared to traditional CSMA/CA based scheduling.

Figures (12)

• Figure 1: Wireshark I/O graph illustrating the operation of TWT.
• Figure 2: Schematic diagram of the testbed.
• Figure 3: An illustration of Multiplication Factor (MF.)
• Figure 4: Throughput variation of TWT schedules with MCS index 7.
• Figure 5: Variation of packet aggregation for different AA and MF 40.