Fair and Efficient Dynamic Traffic Grooming Algorithm for WDM Mesh Networks

Traffic grooming in wavelength division multiplexing networks merges low-speed flows into large capacity pipes so that the bandwidth discrepancy between them will not lead to underutilization of resources. Dynamic grooming deals with requests for wavelength allocation based on a dynamic pattern of arrivals in contrast to the situation of static grooming in which the pattern of arrivals must be previously known. Solutions for dynamic grooming typically involve the construction of an auxiliary graph for deciding on the routing and wavelength assignment. An auxiliary graph can represent the network partially leading to scalable solutions, although blocking is not fairly distributed among calls in such solutions. A novel algorithm is thus proposed for achieving fairness in blocking among the calls. In an off-line step, it defines an appropriate set of OXCs for achieving fairness and efficient use of resources. Results reveal a higher degree of fairness obtained by the use of the novel algorithm than for previously proposed algorithms.