Modeling and performance study of restoration framework in WDM optical networks

Restoration techniques available in literature have not addressed their performance in terms of network performance factors and network costs. However, restoration must consider a number of network performance factors and network costs simultaneously. We evaluate existing restoration models and their performance in an attempt to verify their performance and efficacy based on literature. Our research has found not only inefficiency in some of these methods, but a general incompatibility. Consequently, we propose a restoration framework and examine its performance. The proposed restoration framework consists of seven objective functions and three algorithms. Three algorithms are developed for restorable routing and wavelength assignment. Seven objective functions yield their own objective goals significant to the optimal design of a survivable WDM (Wavelength Division Multiplexing) optical network. Three algorithms are exploited to measure the performance of objective functions. They are DCROS (Deep Conjectural Reinforced local Optimal Search), HRM (Hybrid Restoration Method), and RWWA (Random Walk-based Wavelength Assignment) algorithms. The DCROS algorithm finds a set of disjoint lightpaths for a given objective function. Each of those disjoint lightpaths can be used for a primary lightpath and a backup lightpath. The HRM improves results generated by the DCROS algorithm. The RWWA algorithm assigns wavelengths to primary and backup lightpaths. Numerical results obtained by experimental evaluation of proposed objective functions and algorithms confirm that MTLF (objective function of Minimizing Total Lightpath Flows with multi-objective goals) among objective functions yields multiple near optimal objective goals: minimal wavelengths, minimal wavelength link distance, minimal wavelength mileage costs, even distribution of traffic flows, total and average restoration time of backup lightpaths, and escape from physical topology constraints.