版权所有:内蒙古大学图书馆 技术提供:维普资讯• 智图
内蒙古自治区呼和浩特市赛罕区大学西街235号 邮编: 010021
作者机构:Chongqing Univ Posts & Telecommun Sch Commun & Informat Engn Chongqing 400065 Peoples R China Henan Univ Technol Coll Informat Sci & Engn Zhengzhou 450001 Peoples R China Chongqing Univ Posts & Telecommun Inst Intelligent Commun & Network Secur Chongqing 400065 Peoples R China Henan Univ Technol Minister Educ Key Lab Grain Informat Proc & Control Zhengzhou 450001 Peoples R China Tianjin Univ Coll Intelligence & Comp Sch Comp Sci & Technol Tianjin 300072 Peoples R China
出 版 物:《JOURNAL OF LIGHTWAVE TECHNOLOGY》 (J Lightwave Technol)
年 卷 期:2025年第43卷第5期
页 面:2039-2052页
核心收录:
学科分类:0810[工学-信息与通信工程] 0808[工学-电气工程] 08[工学] 0702[理学-物理学]
基 金:National Natural Science Foundation of China [62025105, 62401197] Natural Science Foundation of Chongqing [CSTB2024NSCQ-MSX0375] China Postdoctoral Science Foundation [2024MD754042] Grain Information Processing Center of Henan University of Technology [KFJJ-2020-111]
主 题:Monitoring Optical switches Location awareness Accuracy Machine learning Integrated optics Costs Training Predictive models Machine learning algorithms All-optical data center networks (DCNs) integer liner program (ILP) optical path failure localization machine learning-based on monitoring trail (mlm-trail)
摘 要:We propose Machine Learning-based Monitoring Trail (mlm-trail) to monitor optical path failures for large model training. Existing monitoring trail (m-trail) method can only localize optical link failures, whereas mlm-trail differs from it by combining with machine learning and providing bidirectional voltage constraint. mlm-trail can provide fast, accurate and integrated failure localization of both optical links and optical switches using only a small number of monitors. Firstly, we construct an input dataset based on the edge relationships of 10000 virtual network topologies (the mappings of optical links and optical switches in all-optical data center networks (DCNs)). Then, monitoring trail under bidirectional voltage constraint is formulated by integer liner program (ILP) to minimize the overall monitoring cost (including monitor and bandwidth costs), and thus construct an output dataset. Finally, we train the learning model based on above dataset using classical and proposed hybrid machine learning models to achieve fast generation of monitoring trails. Based on the constraints of full coverage of monitoring trail and uniqueness of alarms, the generated output results based on machine learning are modified to achieve unambiguous localization for each optical path. Numerical results show that mlm-trail outperforms m-trail in localization speed and scalability, and also outperforms machine learning algorithms in accuracy and cost.