Optimizing enterprise wireless networks through centralization

As the number of wireless devices continues to grow in offices and enterprise environments, Wireless Local Area Networks (WLANs) have emerged as an important part of an enterprise network. Most enterprise WLANs tend to have a centralized architecture, which facilitates management and better design of various control and security functions. In spite of significant progress made in planning, deploying, and managing such enterprise WLANs, radio interference remains a core concern among WLAN users, network administrators, and operations staff alike. This dissertation explores the design and implementation of data and control plane mechanisms that leverage the centralized architecture of enterprise WLAN's to manage interference effectively in such environments. We first explore the design of a centralized data plane for enterprise WLANs and present CENTAUR, a centralized scheduling framework that mitigates interference by intelligently scheduling downlink packets to avoid simultaneous transmissions on conflicting links. Next, we explore control plane mechanisms that manage contention by configuring the operating parameters for the wireless APs. As a part of our efforts into the control plane, we present Model-TPC, a mechanism that facilitates robust and practical transmit power control in enterprise WLANs by determining the set of feasible power levels for different APs in the system. Both CENTAUR and Model-TPC require real-time interference estimates to function efficiently in dynamic wireless environments. Towards that end, we propose PIE, an online interference estimation mechanism for enterprise WLANs that merges traffic reports from wireless APs to generate dynamic interference estimates. Finally, we believe that the mechanisms presented in this dissertation are simple, yet effective and can be used as building blocks for designing more sophisticated tools towards managing interference in enterprise WLANs.