Dimensioning Wavelength-Routed Multi-Fiber WDM Networks with Limited Number of Wavelength Converters

Sparse-partial wavelength conversion in WDM networks has been shown as a very efficient network architecture. We present a network analysis and network dimensioning approach for stochastic path requests based on a reduced load approximation. This model enables us to dimension the number of WDM-systems per link and the number of wavelength converters per node needed to guarantee a maximum blocking probability for a given traffic matrix. The key feature of our approach is that the traffic dynamic is one of the free model parameters. This allows us to model the wide range between static traffic and Poisson traffic. A numerical example shows how the dimensioning result depends on the traffic dynamic.