Dynamic Traffic Allocation in IEEE 802.11be Multi-link WLANs

TL;DR

This paper proposes a dynamic traffic management approach for IEEE 802.11be multi-link WLANs, demonstrating significant improvements in traffic congestion, quality of service, and network coexistence through simulation-based evaluation.

Contribution

It introduces a traffic manager with dynamic allocation policies for MLO, showing enhanced performance over static methods and better coexistence with non-MLO networks.

Findings

Dynamic policies reduce traffic congestion and improve QoS.

MLO enhances network performance and coexistence.

Traffic balancing benefits both MLO and non-MLO networks.

Abstract

The multi-link operation (MLO) is a key feature of the next IEEE 802.11be Extremely High Throughput amendment. Through its adoption, it is expected to enhance users' experience by improving throughput rates and latency. However, potential MLO gains are tied to how traffic is distributed across the multiple radio interfaces. In this paper, we introduce a traffic manager atop MLO, and evaluate different high-level traffic-to-link allocation policies to distribute incoming traffic over the set of enabled interfaces. Following a flow-level approach, we compare both non-dynamic and dynamic traffic balancing policy types. The results show that the use of a dynamic policy, along with MLO, allows to significantly reduce the congestion suffered by traffic flows, enhancing the traffic delivery in all the evaluated scenarios, and in particular improving the quality of service received by video flows. Moreover, we show that the adoption of MLO in future Wi-Fi networks improves also the coexistence with non-MLO networks, which results in performance gains for both MLO and non-MLO networks.