User Tools

Site Tools


Traffic pattern analysis and load balancing in negatively curved networks

Edmond A. Jonckheere
Electrical Engineering—Systems
University of Southern California

Wednesday, April 14th, 2010
EEB 248

Abstract: In the quest to identify a single parameter that would encapsulate such relevant properties of networks as their potential to create congestion, network investigators have for the past decade focused their attention on the Small-World/Scale-Free dichotomy. Here, it is argued that the all encompassing parameter that predicts network behavior is curvature. Curvature indeed anticipates congestion, should the network be negative curved. Curvature indeed anticipates a balanced load, should the network be positively curved. Curvature indeed guarantees successful greedy forwarding of packets in sensor networks, provided the curvature has been uniformized by the Ricci flow.

Back to CommNetS Seminar Page

traffic_pattern_analysis_and_load_balancing_in_negatively_curved_networks.txt · Last modified: 2016/09/01 19:15 (external edit)