Two Layers Of Optical Network Planning Optimization Algorithm

With the rapid development of WDM (wavelength division multiplexing), OXC (optical cross-connect) and OADM (optical add-drop multiplexing) technology, WDM optical network can provide huge bandwidth and has become the core technology of next generation networks. Due to the excellent functions (QoS, TE) together with the implementation in the unified control plane, MPLS technology becomes the best choice to integrate IP and WDM networks, and MPLS over WDM network developed rapidly.The wavelength of WDM optical networks can provide up to more than G bits transmission capacity (for example), but in real applications, the granularity of low-speed demands is much smaller, such as OC-1, OC-3, OC-12(51.84Mb/s,155.52Mb/s,622.08Mb/s). Obviously, to accommodate such kind of low-rate traffic demands (or connections) with one lightpath will lead to inefficient resource utilization. At the same time, it is impossible to establish end-to-end lightpaths for all the traffic demands, due to the limits of the number of wavelengths per fiber and the number of receivers per node. So it is necessary to combine the low-speed demands onto high-speed lightpaths in multi-layer optical networks, this scheme is the so called traffic grooming technology.WDM optical networks transmit large number of traffic demands with its huge bandwidth, but this also brings great challenges. A single failure of any network equipments may affect large amount of demands and cause great loss. So it is necessary to take the fault recovery abilities of optical networks into account. Hence, survivable study has been playing more and more important roles in the multi-layer network design.In this paper, the author focuses on the survivable traffic grooming problem under SRLG (Shared Risk Link Groups) constraints in MPLS over WDM mesh networks. Survivable traffic grooming problem can be described as follows: Given the network configuration, including the physical links and nodes, the number of transceivers and receivers, the number of wavelength in each fiber and the wavelength capacity, survivable traffic grooming can effectively grooming low-speed demands onto high-capacity lightpaths as well as provide protection to demands, and improve the network throughput or reduce the network cost.This paper studies the following two problems: (1) Given enough network resources, the objective is to minimize the total used wavelengths in physical networks. Here the author proposes a novel Link-Path based ILP model and a novel heuristics algorithm named CLIR-MLTG (Cross Layer Information Routing & Multi-layer Traffic Grooming). Compared with the Node-Link based ILP formulation, the proposed Link-Path based ILP model reduced problem size and execution time. Comapared with the traditional heusistic methods, CLIR-MLTG heuristic algorithm avoids adding lightpaths when unnecessary, and then minimizes the total used physical resource. (2) With limited network resources, the objective is to maximize the network throughput as well as maximize the network revenue. In order to solve this problem, the author proposes a cross layer iteration ILP algorithm based on Lagrangian Relaxation (LR), and decomposes the total optimization problem into two sub-problems in MPLS and WDM layer, and get the upper and lower bounds through interactive iteration of the two-layer data, and then estimate the optimized results of the whole optimization problem.