The Monitor And Analysis System Of Online Payment Risk Management Platform

With fast development of mobile network technologies,data is been created at a rapid speed.At the same time,online payment has become the main target for hackers and fraudsters.Starting from 2010,the rate fraud action of online payment is growing at 20 percent per year.Multiple payment companies started to build decision framework and neural net model to detect and control fraud actions.Thus there is need to monitor the daily operation and availability of risk control systems.As the value of data decays with time,the biggest challenge of data analysis is how to process big data volume instantly.So there is need to research and construct a real time computation framework to extract valuable data by processing massive server logs and detect risky transactions and fraud attacks.The thesis researches the feature and mechanism of distributed cluster architecture and emphasizes the importance of stream based real time analysis and monitoring of massive risk data in online payment systems.Then the thesis raised three major challenges of real time data processing: big data acquiring and storage,real time analysis and configurable data computing,storage and query in different payment phase.Finally,the thesis provides corresponding solutions to those challenges.The thesis designs and implements an architecture based on distributed computation framework,stream data processing and openTSDB technology.The system archives massive data collection and storage in big distributed data sources.In pre-process phase,the system leverages Spark framework to implement real time data analysis and develop storage and query mechanism using OpenTSDB database.In frontend module,the system utilize existing Spring MVC framework for monitoring and demonstration.In project experiment,the platform leverages features of distributed data collection framework and manages to collect massive raw data from millions of data servers.Powered with Storm data streaming and dynamic time processing feature of TSDB,the platform can perform real time monitoring and analysis of massive online payment data.The report of the platform proves high efficiency of data processing regarding real time and total throughput.