Design And Practice Of Alarm Filtering Mechanism For Centralized Alarm Monitoring System In Local Network

China Telecom's local telecommunications network with multi-specialty, multi-vendor network management platform for the complexity and heterogeneity of the characteristics made the local network operation and maintenance difficult, and labor and management are also high. The local telecommunications network carries on the maintenance organizational reform, then a set of support system, which is integrated centralized monitoring and maintenance, is urgently needed to establish. It promotes local network centralized monitor maintenance level, changes the passive maintenance to the active maintenance, the maintenance work will be changed from network-oriented to customer-oriented, service-oriented, and improve the customer service quality effectively, enhance enterprises in the modern industry's competitiveness. In order to alarm information centralized monitoring and analysis on multi-specialty and cross-platform , the local network alarm monitoring system which is one of the first choice to achieve a system, it is to achieve the reunification of the user interface, to ensure the concern alarms and to detect, position.Centralized alarm monitoring system in local network needs to collect and process the huge number of alarm data from the exchange network management, transmission network management ,data network management, dynamic environment network management and etc. In order to improve the operating efficiency, the system must have the function of alarm filtering. On the one hand it is to ensure focus on the effective alarm data rapidly and accurately, prevent alarms storm effectively. On the other hand to achieve the alarm information automatically associated with customers, and improve the customer service quality effectively. Then, the key is that designing and implementation effective filtering mechanisms of alarms, It's also one of the core tasks of centralized alarm monitoring system in local network.Firstly, This article elaborated on the related knowledge about alarm filtering, such as the concept of TMN network management, the warning, the knowledge and knowledge base, etc. And elaborated on the common methods of alarm relativity analysis and association rule mining algorithm in detail. Secondly, it treats how to obtain the alarm association rule using data mining technology. On based of that, this article constructs alarm filtering rule knowledge base and knowledge expression of every alarm filtering module and knowledge acquisition and updating methods. Meanwhile, this article designs stratified alarm filtering model and template-based filter alarm scheme, illuminates knowledge base rule matching and the methods to settle the collisions. Thirdly, this article elaborated on the concrete implementation of alarm filtering mechanism and algorithm improved through combining classical Apriori algorithm and time window. Finally, this article demonstrated the application effect of the alarm filtering mechanism and manifests the alarm filtering effect with the warning compression ratio.This article applies data mining and knowledge base technology to alarm filtering mechanism. So the research content and results have fine academic, technical and practical value. The research content's innovation mainly manifests in four aspects. The first is, according to the warning data characteristic and the alarm filtering demand analysis, to define the stratified alarm filtering model; The Second is to design alann association knowledge base and produce alarm filtering module intelligently; The third is to acquire alarm time association rule and apply to practical system through combining classical Aprior algorithm and time window; The fourth is to realize stratified converging alarm data fast, flexible configuration and implement of alarm filtering in usage of multi-filtering technology based on module。...