MU-MIMO-Aware AP Selection for 802.11ac Networks.

Major Wi-Fi Access Point (AP) vendors worldwide seek to provide gigabit wireless connectivity, by densely deploying MU-MIMO capable APs, which can support multiple, concurrent data streams to a group of clients, connected to them. However, MU-MIMO gains can only be achieved if an AP can identify groups of clients with homogenous configurations and orthogonal wireless channels, where concurrent transmissions will not cause inter-client interference. Hence, MU-MIMO performance is fundamentally depending on how the clients are assigned to APs. Our experiments with 802.11ac commodity testbeds show that state-of-the-art client assignment algorithms are MU-MIMO oblivious and limit the MU-MIMO grouping opportunities in realistic settings. In this paper, we design and implement MAPS, an MU-MIMO-Aware AP Selection algorithm that is 802.11-compliant and can boost network's MU-MIMO throughput gains. We verified MAPS' gains over legacy designs via extensive experiments with 802.11ac commodity testbeds.