Optimizing telecommunications networks for efficiency and survivability

Abstract

A survivable network needs to be able to deliver committed quality of service under multi-hour traffic and unexpected or urgent circumstances. In this project I consider various scenarios where network design should consider survivability and efficiency while assuring the quality of service. To minimize the network cost as well as satisfying the capacity requirement, considering the varied traffic distribution during different times of a day, I formulated the multi-period traffic as a linear programming problem and obtained the optimized traffic capacity. I observed the traffic variations in a fully-meshed network simulated by MATLAB and computed the optimized link capacity under situations of both multi-period traffic and link break. Then I created animations to illustrate the mechanisms of two survivable networks, UPSR and BLSR by Adobe Flash and sketched the queue size changes of each node in MATLAB by simplifying them into unidirectional and bidirectional ring networks respectively with Jackson Theorem.