The Research And Application Of Transaction Monitoring Model In Financial System

As the demands of individual users being increased in recent years,many enterprises led by those in the Internet business,have conducted in-depth research to meet customers' demand.With that,the enterprises have begun to invest more in information systems and constructed multi-type systems.With the increasing number of application systems and system-level hardware and software,a comprehensive monitoring framework,consisting of hardware monitoring,computer monitoring,operating system monitoring,network monitoring and open platform monitoring,have been formed through the construction of monitoring platform.Yet in daily operation and maintenance,it is commonly seen that the online transaction is slow with a normal system resource,or the time for daily batch transaction is prolonged.In order to solve the monitoring problem of online transaction and prolonged batch transaction,this paper has constructed a transaction monitoring model in information system based on mass data.Through the gathering of transaction log,transaction data and alarm information,raw data are integrated and transaction level data are thus formed and stored in the database;through the analysis of transaction data by analyzing programs,the transaction response time,success rate,concurrency and batch processing time of transaction level are acquired;through the data analysis of online transaction and batch transaction,the changes of the transaction trends could be got.Combined with the fundamental monitoring data,the reasons(CPU,disk IO or memory)are found to help the operation and maintenance personnel to position the problems and predict the adverse trends to avoid the occurrence of faults.The analysis of transaction monitoring model in this paper includes two parts,online transactions and daily batch transactions.The transaction information in original production system was collected through scripting language,including the transaction messages in transaction logs and the data of transaction process in database.These two types of data were organized into standard data format for analysis.After this,the final result of transaction analysis by analyzing program is gained.Then,a comparison is made between it and the result generated by the monitoring system to obtain the particular program and exact fundamental resource resulting in the system delay,and the method of optimization is thus got.Afterwards,the optimized transactions are monitored again to test the effect of transaction monitoring.