Assessing the costs (and benefits) of dependable public IP service networks

Considering the problem of estimating the financial impact of dependability improvements for commercial IP networks. An economic reliability analysis framework (called ERA) was developed at the university [2003] and applied it to a modern, commercial IP network to address the problem. The ERA framework extends traditional reliability engineering techniques to include financial elements such as: revenues, expenses, and investment periods. Full-path enumeration techniques [2003] were already extended. Markov reward models (MRM) [2003], software reliability growth models (SRGM) [2003] and stochastic reliability models [E.J. Stoker et al., 2003] using the ERA framework. This paper integrates ERA framework to reliability bounding techniques and using it to forecast the expected reliability and expected economic efficiency of a commercial IP data network. The commercial network being analyzed links 21 distinct locations along with 2 trans-oceanic edges to another IP network, each with its own collection of nodes, edges, customers and local topologies. Changes are proposed to this network to improve overall customer dependability, and the impacts of these changes (in terms of network dependability and economic return) are analyzed.