Throughput Stability and Flows in Large-Scale Random Networks

Sergio D. Servetto, Eva TardosThe rate of growth of the maximum stable throughput in large-scale random networks as a function of network size is studied in this thesis. The problem is formulated as one of determining the value of the maximum multicommodity flow on the corresponding random unit-disk graph and shown to be equivalent. In this way, using simple flow techniques and probability tools, a tight bound is derived on the rate of growth of the maximum stable throughput with a fairness constraint. As an application of these techniques, similar bounds are computed for different cases of highly dense wireless networks when directional antennas are being used and the results are compared to the omnidirectional case.National Science Foundatio