Research On Traffic Grooming Algorithm In The Architecture Of IP And Optical Convergence

The IP and optical convergence architecture is the mainstream architecture of the backbone network in the future,and it is also a hotspot of communication research.The diversity of services often makes the bandwidth required by the service not match the "rigid" bandwidth provided by the optical network.It is a great waste for the network to allocate the end-to-end optical channels for each service.Traffic grooming can save network resources by aggregating multiple fine-grained services together and processing them as a single coarse-grained service.In this thesis,we mainly study the traffic grooming problem by reducing the number of guard-bands and transponders.The specific research is as follows:(1)For the problem that the filter in the optical network is not an ideal filter and the spectrum resource is wasted by adding the guard-band between adjacent optical channels,a distance adaptive algorithm for optimizing guard-band(DAA)is proposed.The algorithm combines the distance with the modulation format.The closer the distance,the higher the modulation format that can be used,and the less spectrum resources are required to be carried on the optical.At the same time,the virtualized technology is used to virtualize the established direct link resources,and a feasible grooming strategy is proposed.The virtual direct link established by the algorithm under different physical hop limit is compared with the existing algorithm.The best result is a 11.7%savings in guard-band.(2)In order to improve the utilization of transponders in the network,an integer linear programming(ILP)model with the goal of minimizing the number of transponders is proposed.The key to model establishment is to convert the transponder capacity into spectrum resources in the optical network.Marking the last bit slot of within the capacity and converting the tag into the number of transponders for modeling.The model is compared with a heuristic algorithm for maximizing electrical-grooming and a non-grooming algorithm.The linear programming model can achieve a global optimal solution in the case of a small network size.(3)Introducing algorithm simulation platform for the IP + optical fusion path computation and resource scheduling based on the proposed algorithms,and the functional design module of the platform is also proposed.