Based On A Bank Of Association Rules In The System Managed Communications Web Alert Correlation Analysis

The modern commercial bank network architecture is getting more and morecomplex, this is because bank organization scale expands and application concentrationdegree unceasingly constantly enhances to decide. Because the time of networkconstruction project is different, therefore each node is built by the network equipmentof different manufacturers. The massive warning information that the networkmaintenance personnel obtain often are the derivation warning information that iscaused by some node breakdown. These above factors gave modern commercial bank’snetwork management to face more and more difficulties. This article uses the networkenvironment of some large state commercial bank Sichuan Branch as the analysisbackground, with this bank from May,2011to May2012, the network warningmessage data as the object of study, will seek for the current network management tocome across the issues and come up with the solutions.This article describes the data mining association rules correlation analysis theoryof basic knowledge and related concepts, and how the practices of the theory for dataanalysis step. Collected during the period of the bank, a large number of networkdevices take the initiative to send or tools collected by syslog log trap alarms andequipment performance data and other information. Bank’s high availability networkrunning status, fault and other information will be hidden in these vast amounts ofhistorical data. Noise reduction through network alarm information collected, analyzeand organize, generated several adapt association rule analysis method based on theassociation rule mining knowledge, are related compression rules, filtering rules,suppression rules, model rules and special rules analysis method. And using the analysismethod of the association rules have been processed, the filtered information datamining, access to a series of useful information for network management data. Finally,details how the preliminary study of the analysis method used in the existing bankingnetwork management system, and the application of the results based on associationrules analysis method presented.This work proposes the analysis and correlation analysis method based on the history of the association rules network log data and performance data through thealarm association rules analysis method applied to the alarm monitoring system runningactual results show that the association rules method to achieve the expected effect. Thisshows that the research topics NMS in the bank can be a good application, while thesubject of the research content also has a strong theoretical value.