Sdh Network Performance Optimization Algorithm

With the rapid development of information technology, people's demands of multimedia information such as graphics, audio and video increase in a highly speed. How to satisfy the need for huge network bandwidth is an important problem. For this reason, OTN (Optical Transport Network) with huge bandwidth develops rapidly. How to plan OTN in an economic and effective way is a hot spot in research field. As a multi-layer network, OTN includes different technologies on different network layers, e.g., SDH (Synchronous Digital Hierarchy). In OTN network planning, how to optimize the performance of SDH network is a priority in network planning.In this paper, the author studies three cases in SDH network planning: the resource allocation, path recovery and timeslot assignment.In chapter 2, the author studies how to allocate resources and reduce congestion while optimizing resource utilization and improve network performance. Author proposes a dual iterative algorithm and a link weight setting methods. The link weight setting method can avoid the appearance of link congestion as possible when routing. The dual iterative algorithm optimizes the performance of the network by global iteration for all demands and local adjustment for some demands.In chapter 3, the author studies how to recovery the current paths to their original ones without the interruption of demands. It is often an engineering problem. Author proposes a concept of a conflict set, and an intermediate state transition method, the maximum possible to achieve non-disruptive path to recovery.In chapter 4, the authors studied the timeslot assignment in SDH network. The timeslot assignment algorithm the author proposed can deal with almost all dynamic demands in mesh networks or ring networks, such as special protection demands, shared protection demands, non-protection demands. In addition, the algorithm can deal with demands which have constrains. The algorithm can reduce the resource fragmentation and improve resource utilization results.In this dissertation, simulation results indicate the proposed algorithms performance well.