Research And Implementation Of Algorithmic IT Operation And Operation Equipment Relation Network System Based On Temporal Association Rules

With the advent of Big Data,the scale of business in enterprise continues to expand,lead to the increment of complexity of business systems.With the emergence of new technologies such as distributed and virtual machines,which aim at ensuring data security and operating business systems stably.But the IT Operation has to run through the entire system because of the cohesion of the distributed system,distributed system will also generate a large amount of operation data in every absolute time.So how to use these data to diagnose the operating status for the entire system has become a subject in Algorithmic IT Operation(AIOps).In this paper,based on operation equipment index data and alarm data in power grid system,using Self Organizing Mapping Network,One Class Support Vector Machine,the combination of negative samples clustering index and alarm data realizes the association rules deduction between the finite state machine and the alarm.The main content of this article is as follows:1.Based on the AIOps and Temporal Series Association Rules(TSAR),a literature review is conducted,which mainly studies the main functions,application scenarios and framework in AIOps;TSAR previous research and problems.Combined with actual application environment analyzes the requirements of the TSAR system and the main problems in system design section.Finally proposes system’s technical development route and framework.2.The outline design of Intelligent operation equipment relation network system based on TSAR,combined with the survey results,put forward the functional requirements analysis of the entire system,including,Overall structure diagram,Overall operation flowchart and performance index requirements.Basic functional modules are given including: Data Pre-processing module,Maintenance Equipment Finite State Machine Mapping module and Single Classification Alarm Detection and Clustering Alarm Association Rules Deduction module.3.For Maintenance Equipment Finite State Machine Mapping module and Single Classification Alarm Detection and Clustering Alarm Association Rules Deduction module,the system core module is designed in details from two parts:Data Structure and Storage and Algorithm Execution process.Clarify the data storage and execution process in each core module.Propose solutions to problems that may occur during the execution of the module at the system level.4.Research the core algorithms which is used in the system,and test the feasibility by Test Data and evaluate the algorithm objectively.The key technologies are: Self Organizing Mapping,One Class Support Vector Machine and Alarm Association Rules.Algorithms are tested using actual data sets,and compared the result with control groups.Results show that the entire system has reached the expected goal in algorithm design,and has strong practical decision-making significance for the association rules.