Path computation elements （PCEs） are employed to compute end-to-end paths across multi-domain optical networks due to the advantages of powerful computation capability. However, PCEs＇ location selection is still an...Path computation elements （PCEs） are employed to compute end-to-end paths across multi-domain optical networks due to the advantages of powerful computation capability. However, PCEs＇ location selection is still an open problem which is closely related to the communication overhead. This paper mainly focuses on the problem of PCEs＇ location selection to minimize the overall communication overhead in the control plane. The problem is formulated as a quadratic integer programming （QIP） model, and an optimal decision rule is gained from the solution of the QIP model. Then based on the decision rule, a distributed heuristic algorithm is proposed for dynamic network scenario. Simulation results demonstrate the benefit and the effectiveness of our proposed approach by comparing it with random selection policy.展开更多
基金supported by the National Basic Research Program of China (2010CB328202, 2010CB328204, and 2012CB315604)the Hi-Tech Research and Development Program of China (2012AA011302)+3 种基金the Beijing Nova Program (2011065)the RFDP Project (20120005120019)the Fundamental Research Funds for the Central Universities (2013RC1201)the Fund of State Key Laboratory of Information Photonics and Optical Communications (BUPT)
文摘Path computation elements （PCEs） are employed to compute end-to-end paths across multi-domain optical networks due to the advantages of powerful computation capability. However, PCEs＇ location selection is still an open problem which is closely related to the communication overhead. This paper mainly focuses on the problem of PCEs＇ location selection to minimize the overall communication overhead in the control plane. The problem is formulated as a quadratic integer programming （QIP） model, and an optimal decision rule is gained from the solution of the QIP model. Then based on the decision rule, a distributed heuristic algorithm is proposed for dynamic network scenario. Simulation results demonstrate the benefit and the effectiveness of our proposed approach by comparing it with random selection policy.
