Cross-Layer Techniques for Wireless Local Area Networks

In this dissertation, we examine important aspects of infrastructure IEEE 802.11 Wireless Local Area Networks (WLANs) and identify issues that can affect their performance. Reviewing the state of the art, we observe that numerous research efforts have proposed diverse solutions with several limitations that impede their deployment in existing WLANs. Moreover, users have ever-increasing expectations of availability, reliability, instantaneous response and security from their wireless connections.
Motivated by these challenges, we design and implement novel but practical solutions that address open issues affecting the performance of IEEE 802.11 WLANs. We adopt an Access Point (AP)-based approach, which does not require any modification in the clients. We focus on the following aspects of WLANs: client mobility, channel management, and quality of service, and explore three different scenarios for the most common deployments: an enterprise, a city (urban area), and a personal residence (home). To provide a common basis for practical implementation of new 802.11 solutions, we present a Smart AP model, inspired by self-management techniques.
The main contributions of this thesis are:

1. We develop a seamless mobility solution for Voice over IP (VoIP) services in Enterprise WLANs, called Multichannel Virtual Access Point (mVAP), which requires no client modifications and is compatible with current devices. We implement and evaluate mVAP using commodity 802.11 hardware, and achieve transparent mobility without interruption or degradation of ongoing communications.
2. We investigate the feasibility of harnessing the existing WiFi coverage in urban areas for mobile Internet access, through trace-based simulations using real data collected by mobile phones. The results show that the WiFi coverage is large and the connectivity it offers can be effectively exploited. We identify open issues for the actual deployment of such a citywide WiFi network and the applications that could benefit from it.
3. We propose an adaptive traffic-aware channel selection mechanism for Home WLANs, that uses the time-varying traffic load for interference estimation. We implement this solution using commodity 802.11 hardware and experimentally evaluate it: the network performance is drastically improved by constantly picking the channel with the least interference.