Uplink Multi-User (MU) MIMO transmissions allow clients to simultaneously transmit independent data streams to the Access Point (AP), effectively multiplying the capacity of the wireless channel for uplink access. The IEEE 802.11ax amendment defines the Triggered Uplink Access (TUA) mechanism as the only way to initiate an uplink MU-MIMO transmission in Wi-Fi, which enables an AP to start and time synchronize simultaneous uplink MU transmissions by broadcasting a trigger frame containing the resource allocation information.
Unfortunately, the aforementioned procedure introduces challenges in the design and implementation of medium access control policies for 802.11 wireless networks, including contention and scheduling problems. In this thesis we present the following three contributions. First, we design, implement, and validate PERFORM, a novel experimental platform for implementation and evaluation of WLAN MAC policies. This platform is the first end-to-end integrated system that carries real-time network traffic and, at the same time, allows for the flexible prototyping and evaluation of advanced WLAN MAC policies, including MU-MIMO, TUA, and buffer status reports. Second, we experimentally study the role of real application traffic on the performance of TUA.
While TUA gains for fully backlogged traffic are well established, we show that bursty closed-loop traffic radically transforms performance.
Last, we introduce Client-side Access Manipulation (CAM) as a mechanism that enables clients to dynamically adapt their channel access priority without explicitly coordinating state with the AP to realize an efficient uplink multi-user WLAN. Our results start by presenting the PERFORM platform validation, which verifies that the platform can achieve the necessary performance targets. Then, using PERFORM, we find that TUA significantly reduces file transfer latency compared to legacy single-user uplink, but unfortunately the standardized method for low-overhead backlog reporting leaves substantial benefits unrealized. Last, we show that CAM achieves gains in throughput and up to 65% reduction in average latency. Moreover, with the use of the standard’s defined access adaptation mechanism on the same scenarios, the aggregate throughput decreases, and the average latency increases sharply.