AP-initiated Multi-User Transmissions in IEEE 802.11ax WLANs

Abstract: Next-generation 802.11ax WLANs will make extensive use of multi-user communications in both downlink (DL) and uplink (UL) directions to achieve high and efficient spectrum utilization in scenarios with many user stations per access point. It will become possible with the support of multi-user (MU) multiple input, multiple output (MIMO) and orthogonal frequency division multiple access (OFDMA) transmissions. In this paper, we first overview the novel characteristics introduced by IEEE 802.11ax to implement AP-initiated OFDMA and MU-MIMO transmissions in both downlink and uplink directions. Namely, we describe the changes made at the physical layer and at the medium access control layer to support OFDMA, the use of \emph{trigger frames} to schedule uplink multi-user transmissions, and the new \emph{multi-user RTS/CTS mechanism} to protect large multi-user transmissions from collisions. Then, in order to study the achievable throughput of an 802.11ax network, we use both mathematical analysis and simulations to numerically quantify the benefits of MU transmissions and the impact of 802.11ax overheads on the WLAN saturation throughput. Results show the advantages of MU transmissions in scenarios with many user stations, also providing some novel insights on the conditions in which 802.11ax WLANs are able to maximize their performance, such as the existence of an optimal number of active user stations in terms of throughput, or the need to provide strict prioritization to AP-initiated MU transmissions to avoid collisions with user stations.

• The paper introduces AP-initiated MU transmissions in 802.11ax that combine MU-MIMO and OFDMA to boost throughput in dense networks.
• It adapts Bianchi’s model to evaluate saturation throughput, highlighting methods for efficient CSI feedback and packet aggregation.
• The study demonstrates that optimized MU scheduling in 802.11ax outperforms 802.11ac, particularly improving downlink performance in high-density scenarios.

AP-initiated Multi-User Transmissions in IEEE 802.11ax WLANs

Introduction

The IEEE 802.11ax WLAN amendment represents a significant evolution in wireless network technology, primarily targeting high-efficiency (HE) scenarios characterized by a dense distribution of user stations (STAs) and access points (APs). It builds upon the legacy IEEE 802.11ac specification, extending MU transmission capabilities by incorporating both UL MU-MIMO and OFDMA transmissions. This paper provides an in-depth analysis of these advancements and investigates the achievable throughput in IEEE 802.11ax WLANs, emphasizing conditions that maximize performance.

IEEE 802.11ax: MU Transmissions in WLANs

Physical Layer Enhancements

IEEE 802.11ax maintains an OFDM-based PHY layer, projecting an increase in channel efficiency primarily via enhanced sub-carrier structures. It extends the OFDM symbol and Guard Interval (GI) durations while implementing 1024-QAM modulation, a significant step up from IEEE 802.11ac's maximum of 256-QAM. These enhancements aim to increase robustness against channel interference and improve data rates, achieving higher aggregate throughput, while maintaining optional transmission flexibility with variable channel widths.

MAC Layer Innovations

The paper discusses the integration of multi-user communication at the MAC layer, highlighting new control frames like Trigger-based (TB) UL MU transmissions and the MU-RTS/CTS procedure. These mechanisms aim to streamline the medium access process and better manage Quality of Service (QoS) requirements in high-density settings. They promise to enhance resource allocation efficiency and spatial utilization.

Data Transmission Strategies

The paper outlines the three MU transmission techniques in IEEE 802.11ax: MU-MIMO, MU-OFDMA, and their combined format for concurrent utilization of spatial and frequency resources. The AP's role in managing these transmissions is emphasized, with new overhead considerations laid out, including the methods for Channel State Information (CSI) and Buffer Status Information (BSI) acquisition.

\includegraphics[width=0.9\textwidth]{FiguresInitial/Figure1.eps}

Analytical Model for Saturation Throughput

Building upon Bianchi's well-regarded IEEE 802.11 DCF model, the paper adapts this model to incorporate the MU-MIMO and OFDMA features of IEEE 802.11ax. This involves defining the achievable throughput for both downlink (DL) and uplink (UL) under saturation conditions, deriving the probabilities of successful transmissions and collisions for both SU and MU operations, and quantifying overhead expenses introduced due to channel sounding and MU transmission scheduling.

Key Findings and Practical Implications

The analysis reveals several insights:

1. Efficiency Gains Through MU Transmissions: The introduction of large MU transmissions and packet aggregation is crucial to achieving higher efficiencies in dense environments where one AP manages numerous user stations.
2. Important Numerics:
   • MU considering up to 64 stations shows significant throughput increases, particularly in DL.
   • The efficiency of CSI feedback is highlighted, revealing the reduced overheads when stations transmit CSI reports simultaneously using UL MU features.
3. Trade-Offs in Performance: While enabling the AP to schedule large DL transmissions increases network efficiency, this comes with the need for periodic CSI collection, which can hinder performance due to inherent overheads and potential outdated channel estimations.
4. Comparative Analysis with IEEE 802.11ac: In scenarios with few user stations, IEEE 802.11ax shows enhanced throughput even when compared with IEEE 802.11ac, fundamentally due to the increment in the maximum number of frames that can be aggregated in a single A-MPDU and larger MU transmissions.
5. Configuration Impact: The IEEE 802.11ax performance is sensitive to parameters like channel width, user station numbers, spatial streams, and A-MPDU sizes. Misalignment or suboptimal settings could significantly degrade throughput, particularly in the DL direction.

Conclusion

This paper offers a comprehensive investigation into the performance implications of AP-initiated MU transmissions as defined in the IEEE 802.11ax draft. The results underscore the significance of maximizing MU transmission under high-density scenarios for optimal throughput. It also reveals the essential nature of efficiently managing control overhead like CSI and BSI acquisition to leverage the full potential of IEEE 802.11ax capabilities. Future avenues of exploration should include non-saturation traffic models, further channel condition analysis, and the development of more efficient scheduling algorithms to enhance performance in dense WLAN environments.

Given the development phase of the IEEE 802.11ax amendment, continued research is required to refine both theoretical models and practical applications to fully utilize MU transmissions, ensuring high efficiency in next-generation WLANs.