Design And Implementation Of Business Tracking System For Electronic Commerce Operation Platform

With the development of electronic commerce, the electronic commerce system has been becoming more and more complex. Now e-commerce not only needs to support to a variety of merchandise display, transaction and distribution, but also needs to deal with the huge hits with promotion of e-commerce. In order to meet these needs, an e-commerce system needs to be supported by many operating systems. A successful shopping needs many operating systems cooperating with each other to complete. In this time any system problem will lead to the failure of shopping, and locating and investigation of the operating system problems will bring great difficulties.In order to locate and investigate problems more efficiently and faster in the complex operation platform, this thesis analyzes the demand for business tracking system for e-commerce, designs and implements business tracking system for e-commerce operation platform. Through the log data acquisition and processing, the system is able to reproduce the entire business execution process and progress, facilitate the positioning problems and failure investigation. Meanwhile, by comparing the data clustering results and historical data, the system realizes the business risk early warning function. The main works of this thesis are as follows:(1) The design scheme of business tracking system for the e-commerce operation platform is put forward. According to the input and output of the system, the requirement of the system is analyzed, the main function of the system is combed, and the overall structure of the business tracking system and the database table structure are designed.(2) The method of log data acquisition and processing in the business tracking system is put forward, including the log reading, filtering, splitting, processing and saving solutions. Among them, the Sunday that string matching algorithm is used to filter the log, the bloom filter is used to delete the duplication log, the method that subsection generates a unique serial number is used to generate the database serial number of the system, etc..(3) The method of tracing and analyzing e-commerce business is put forward, including the data query, sorting, clustering and early warning solutions. Among them, the concurrent queries and sort is used to track log, the clustering algorithm is used to data clustering, and the comparing historical data with data clustering is used to the business warning, etc..(4) The business tracking system for e-commerce is implemented and tested. The system has been deployed and used in the real scene of the enterprise.