Research On Traffic Grooming Of WDM Optical Networks

Wavelength Division Multiplexing(WDM) technology will be the core technology of next generation backbone networks. Each wavelength in WDM can be operated at very high speed, such as OC-48, OC-192 and OC-768(the corresponding rates are 2.5Gb/s,10Gb/s and 40Gb/s). However, in operational networks, the main connection requests are low-rate traffic streams, e.g.,OC-1,OC-3 and OC-12(the corresponding rates are 51.84Mb/s,155.52Mb/s and 622.08Mb/s,respectively).To accommodate such kind of low-rate connection requests with one lightpath will lead to inefficient utilization of network resources. So it is necessary to adopt traffic grooming techniques to pack multiple low rate connections onto high rate lightpaths, that is, by using multiplexing, de-multiplexing and switching techniques to increase the utilization of wavelengths and tranceivers. In WDM networks, the physical topology is made up of the fiber links which are connected by OXCs. The virtual topology refers to all the lightpahts between the nodes, and the connections are routed and switched on the virtual topology. So the traffic grooming problem in WDM networks can be divided into two layer routing problems.Traffic grooming in WDM networks is studied in this paper, and the focus is the on the mesh networks, which covers the following: multi hop traffic grooming and blocking performance analysis, grooming nodes placement in sparse networks and rerouting techniques in dynamic traffic grooming.Firstly, as a theoretical analysis and modeling of multi hop traffic grooming, blocking performance is investigated based on fixed order alternate paths. Two grooming strategies, FOMH and FO-HC-MH are analyzed. The performance is studied under three grooming policies, namely exhaustive sequential, limited-hop sequential and load sharing policies. The FO-HC-MH strategy presented here has the property of utilizing the tranceivers more efficiently. This savings comes from the restriction of hops of light paths can tranverse at the plan phase of the network.Secondly, the placement of grooming nodes in sparse grooming networks under static connection requests is explored and a strategy based on the induced node degrees is presented. The performance of this strategy is also investigated. As the network size growing large, finding solution of ILP formulation becomes more difficult. A heuristic algorithm is also proposed. The simulation results indicate that choosing and placing grooming nodes properly can even achieve the performance of full grooming networks at a much lower cost.As the load grows higher, the rerouting techniques are investigated in the fourth chapter of this paper. The purpose of rerouting is to carry more connections in the network, utilize the resources more efficiently and lower the block probability. According to the network layers operated, two rerouting schemes are explored, light path level rerouting and connection level rerouting respectively.