Performance Of Wavelength-Routed Optical Networks With Wavelength Conversions

From the emergence of the concept of the all-optical networks at 1990, to the widely commercial use of Dense Wavelength-Division Multiplexing, the rapid advance of optical communication makes Optical Transport Networks (OTN) the only candidate for the backbone network of Next Generation Internet. As a well-developed OTN technology, Wavelength Routed Optical Networks (WRONs) have attracted more and more concerns.As the important components in WRONs, wavelength converters play an indispensable role in eliminating the wavelength continuity constraint, improving the reusability of wavelengths, enhancing the flexibility of the networks, keeping compatibility when interconnecting the equipments provided by multi-vendors. It is of great practical and theoretical significance to research the performance influence of the wavelength converters in WRONs.In this thesis, we focus on the performance issue of WRONs with wavelength conversions, especially the limited-range wavelength conversions. The main works and innovations are as follows. Partial wavelength conversions, which configures several full-range wavelength converters only at some nodes in WRONs, provide a more commercial wavelength conversions configuration. We present an approximate universal analytical technique (E-RLA), based on link independent assumption and RLA approach, for the blocking performance analysis of all-optical networks with partial wavelength conversion, fixed routing and random wavelength assignment. The proposed technique can be applied to arbitrary networks with arbitrary wavelength conversion configuration and arbitrary traffic mode. We further propose the simplified analytical approach (S-RLA) on the basis of the E-RLA technology, to reduce the high computation complexity of E-RLA without loss of the precision of E-RLA. Based on current Max-RWA model, the extended Max-RWA model is presented by introducing limited-range wavelength conversions. Considering signal distortion introduced by wavelength conversions, we propose a new optimization objective, which is to establish the maximum number of connections with the least number of wavelength conversions. The extended Max-RWAproblem is formulated as two Integer Linear Programming (ILP) from the different viewpoint, and then solved using genetic algorithm and tabu-search algorithm respectively. A dynamic routing and wavelength assignment algorithm (LCWA) for limited-range wavelength conversions is proposed. For each randomly arrived request, the correlation of different wavelengths on different, links in all its possible routes is described by wavelength weight in LCWA algorithm. Based on wavelength weight, we try to assign wavelengths on each candidate route using shortest-path algorithm and choose the best as its route and wavelength assignment scheme. We also presented the methods to apply the LCWA algorithm in centralized and distributed network environment respectively. The optimal multicast wavelength assignment problem in WRONs with limited-range wavelength conversions is addressed. Given the multicast request, the optimal object of the problem is to minimize total number of conversions required. In wavelength assignment, we should consider the requirements that the number of needed converters along the path from multicast source node to any member cannot be larger than specified value because of the signal distortion introduced by wavelength converters. At last, we present several dynamic programming to get the exact optimal wavelength assignment scheme under session blocking policy and destination blocking policy.