Research On Service-Oriented Intelligent Fault Location In Optical Networks

With the rapid development of intelligent optical networks,the optical network topology is constantly evolving towards complexity,carrying more user services,and more physical devices are connected to it.Therefore,the corresponding optical network operation and maintenance management costs are also increasing.Among them,network fault management is the top priority.A single fault may even cause hundreds or even thousands of alarms.How to locate faults effectively based on alarm log and improve the stability of the entire intelligent optical network is a problem worth solving immediately.This thesis starts from massive alarm data,considers the service characteristics of the intelligent optical network,and studies the service-based intelligent fault location of the optical network.Some innovative and effective results have been achieved,which are mainly divided into the following aspects:First,a business-oriented mining method for alarm association rules is proposed using data mining algorithms.This method can make full use of massive alarm data and obtain alarm association rules with certain support and confidence.This method mines the alarm rules according to the traffic flow direction,which conforms to the multi-plane and multi-dimensional characteristics of the intelligent optical network,and the generated association rules have a high degree of reliability,which can support the subsequent traffic-based fault location algorithm.Second,based on the service characteristics of the intelligent optical network,a fault location algorithm based on the intelligent optical network service is proposed.This algorithm is based on the service routing data,makes full use of the service topology flow of the optical network,and locates the fault location through the root cause alarm.The root cause alarm spanning tree is generated from the two perspectives of alarm association rules and alarm fault types,and finally the root cause alarm is determined,which solves the problem that the existing alarm analysis algorithm is not related to the traffic flow and improves the accuracy of fault location.Third,from the perspective of automation and intelligence of intelligent optical network operation and maintenance,an intelligent optical network alarm traceability system is designed and implemented.This system can remotely collect data from the optical network management system,perform traffics and topology analysis,and output the root cause of the alarm.spanning tree.