M.S. Thesis Defense - Aditya V. Dhananjay

Date: 18 Apr, 2008
Time: 11:30 AM
Location: 217 Atanasoff Hall
Topic: Load Estimation in IEEE 802.11 wireless networks
Major Professor(s): Lu Ruan


Abstract:

Careful management of resources can significantly improve the overall performance of the wireless network. Examples of these management techniques include association control, channel selection and power control. The efficacy of these functions relies on a meaningful load estimate. This is the motivation behind our work: to design metrics and techniques for load estimation, in the context of network management functions.

We have designed client-aware and client-unaware load estimation schemes for 802.11 infrastructure networks. Through simulations, we observe their behavior to changing network usage patterns, and argue why they perform better than traditional load metrics.

Through testbed experiments, we have validated our load estimation techniques, by using association control as the target application. We modified the client unaware downlink load estimation technique, to generate a unified load estimate that captures an intuitive notion of the amount of contention in the network. We also demonstrate the performance of our association control algorithm, by showing how that overall network performance is significantly increased by application of our technique.