A Minimization Cost Heuristic Approach for Traffic Grooming in IP-over-WDM Networks

Traffic grooming can be used in IP-over-WDM networks to efficiently utilize the bandwidth of wavelength channels. However, providing grooming capability to network nodes increases the network cost that operators must sustain when compared with a non-grooming network. In this paper a minimization cost heuristic approach is presented that seeks for the cheapest IP-over-WDM grooming network. This approach is manageable, allowing an easy study of large networks, and it is generic since it can be applied to OXCs with different switching granularities and/or number of ports. The proposed approach is able to reduce the network grooming cost by efficiently using the induced bandwidth that results from the OXC grooming granularity being coarser than the bandwidth granularity of IP LSP requests. Adaptive and non-adaptive weight functions, that determine the best grooming operation for an arriving IP LSP request, are also proposed and evaluated.