Research On Multicast Routing And Wavelength Assignment Algorithms In Intelligent Optical Network

Routing and Wavelength Assignment(RWA) is a critical issue in Intelligent Optical Networks(ION). With the increasing demands from various multicast applications such as IPTV, multimedia conferencing, and their bandwidth-hungry requirements for differentiated QoS, the Multicast RWA(MRWA) has become an important and challenging topic in optical networks. This thesis gives a comprehensive research on MRWA and its support for distinction services in sparse splitting WDM networks. Theoretical analysis and numerical simulation methods are implemented to verify the performance of improved algorithms.Firstly, this thesis introduces the fundamental principles and some key techniques of ION,focusing on existing MRWA problems solutions in ION. Considering wavelength continuity and sparse splitting constraints, an improved Edge Removal Light-Hierarchy based algorithms(Smart Edge Removal Light-Hierarchy for MRWA, SERLH-MRWA) is proposed, both the cost and remaining resources of the network are taken into consideration, by which the improved algorithm can choose suitable paths depending on network traffic load conditions dynamically. Furthermore,an Improved Wavelength Reservation Based Priority Scheme(IRES) is proposed, which can update reservation wavelengths for each level service and consisting proportions with a predetermined period of time. Combing the wavelength assignment algorithm and SERLH which introduce remaining wavelength ratio of each level together, an improved IRES-MRWA algorithms is proposed to support differentiated service, i.e. the network resource utilization for each level can be taken into consideration while routing. Theoretical analysis and numerical simulations are implemented in different network topology, experimental results showed that SERLH algorithm has better blocking probability than ERLH algorithm in most traffic load scenarios. IRES-MRWA algorithm can not only support differentiated service, but also has good performance such as blocking probability and resource utilization efficiency.