Algorithm of Traffic Matrix Estimation in IP Backbone Networks

Traffic matrix estimation is significantly difficult because it is a highly ill-posed problem. How to overcome the ill-posed nature of this problem is the main challenge faced at present. This paper studies the large-scale IP (Internet Protocol) traffic matrix estimation and uses backpropagation neural network to capture the characteristics of traffic matrix. By describing traffic matrix estimation into an optimal process under the constraints, ill-posed nature of this problem can successfully be avoided. Simulation results show that the estimation algorithm based on backpropagation neural network brings the evident performance improvement.