Integrated optimization of aggregation and core for varying NGN architectures

Abstract

The tremendous increase in network traffic compared to relatively constant per customer revenues forces network operators to change both their aggregation and core networks. The key questions in this context are how cost-effective carrier-grade Ethernet transport abilities, high-capacity WDM transport and scalable, high-capacity IP routers influence the structure and the architecture of next generation networks: What is the right size of an IP backbone? How should aggregation be organized? How (de-)  centralized should services like video-on-demand be provided and how (de-)  centralized should functionalities such as accounting be implemented?

In this paper, we present a model-based optimization approach using mixed-integer programs at its core to answer these questions. We illustrate our general solution approach, describe the mathematical programming models used to solve the arising subproblems, and discuss how to express common network architecture variants within these models. Computational studies on nation-wide reference networks with more than 1800 locations, which have been developed within the German research project 100GET, show the effectiveness of our approach and its usefulness in evaluating alternative network architectures and network expansion strategies.