The Design And Implementation Of The Storm-based Real-time Marketing Activity Monitoring And Anti-fraud Platform

In recent years, marketing activities based on the Internet are becoming more and more,which get a good effect. Meanwhile, a variety of fraudulent activities also start occurring widely,causing huge losses to the relevant enterprises, so we have to take appropriate measures to identify fraud effectively.The traditional marketing activity monitoring and anti-fraud platform always relies on relational database management system, modeling offline and other techniques to monitor the marketing activity and recognize fraud. However, with the number of participants rising rapidly,and the requirement of real-time becoming higher and higher, the traditional platform can not supply the demand of nowdays.Based on the actual project of Guotai Junan Securities Inc, taking Kafka as a message queuing system, in using of Storm combining with cache Redis to calculate, presenting the result with Web Socket and HTML5, the paper constructs a new marketing activity monitoring and antifraud platform successfully. The new platform realizes distributed data colleting, real-time data transmitting, calculation of streaming data, cache of intermediate result, real-time pushing and result presentation, greatly improving the performance and scalability. It achieves millisecond level response and thuoughput of gigabytes each day, has solved the problems of the traditional platform.