A Scheduler for Real-Time Service in Wi-Fi 8 Multi-AP Networks With Parameterized Spatial Reuse

TL;DR

The paper tackles delivering real-time applications over dense Wi‑Fi 8 networks by exploiting Parameterized Spatial Reuse (PSR) in a Multi‑AP coordinated setting. It presents a PSR‑aware scheduling framework that reorders non‑RTA uplink transmissions to maximize PSR opportunities for RTA STAs, formalized as a lexicographic optimization of the maximum consecutive PSR‑unavailable gaps and solved via a greedy algorithm with complexity $\mathcal{O}(M\times N^2)$. The approach yields significant reductions in the $0.999$‑quantile of delay for RTA traffic without degrading non‑RTA throughput or fairness, demonstrated through extensive ns‑3 simulations, and scales to multiple overlapping BSSs. Overall, the work provides a practical, scalable scheduling solution that leverages Wi‑Fi 8 features to meet stringent RTAs requirements in dense deployments.

Abstract

Real-time applications (RTAs) require low delays and impose a significant challenge to Wi-Fi. In Wi-Fi, high delays are often caused by waiting for the channel to become idle. This problem can be solved with Parameterized Spatial Reuse (PSR), which allows a station (STA) to transmit its frame with reduced power simultaneously with a triggered uplink transmission in an overlapping network. The PSR opportunity depends on the pathloss between involved STAs, so the same transmission may allow PSR for one STA but not for another one. Thus, to satisfy tight delay constraints in dense overlapping networks, access points (APs) in the same area shall often allow PSR for every STA with RTA traffic. This letter proposes a fast scheduler enabling frequent PSR transmissions for RTA traffic. The scheduler uses Multi-AP coordination, the feature of upcoming Wi-Fi 8. With simulations, we show that it almost halves the delay for RTA traffic and does not deteriorate the quality of service for other traffic compared with an airtime fairness scheduler.

Figures (2)

• Figure 1: Scheme of the scenario.
• Figure 2: The QoS metric. (a) 0.999-quantile of delay. (b) Packet loss ratio. (c) Average throughput. (d) Jain's fairness index of throughput.