The Design And Implementation Of OneLink Traffic Control System

AliExpress(AE)is a cross-border e-commerce platform owned by Alibaba.The AE traffic team mainly introduces website traffic to the AE e-commerce platform by computational advertising.Computational advertising is a science of advertising mar-keting based on internet practice,which studies how to use computational methods to optimize each link of advertising activities.Website traffic is used to describe the number of users visiting a website and the number of pages visited by users.Different from the free traffic generated by users visiting the website spontaneously,the website can also get sufficient and high-value paid traffic by computational advertisements.In the entire life cycle of internet computational advertising,traffic acceptance,that is,the page access and content rendering after users' visits,is an important work in guiding the transformation of advertising,which needs to rely on close cooperation between advertising platform and promoters.Due to natural factors such as language,country and time difference,it is difficult for advertisers,advertising platforms and promoters in the international cross-border e-commerce traffic business to conduct effective com-munication and collaboration,which will lead to lots of invalid and inefficient page visits in the link of paid traffic,and thus fail to increase conversion rate.In this thesis,to solve the above problems,the OneLink traffic control system is designed to intervene in paid traffic acceptance between advertiser's advertising and the buyer's final visit to landing page.The OneLink system provides support for users to customize traffic intervention schemes by their needs and recommended by algorithms,including the intervention on attributes of buyer groups and traffic attributes.When a buyer clicks on an advertisement,the traffic passing through the OneLink system would be matched,rewritten,redirected,and finally presents the actual landing page to the buyer.OneLink system uses Spring Boot framework and AOP(Aspect Oriented Pro-gramming)techniques based on the Spring Boot to implement the business logic,us-ing middlewares like the message queue,service governance framework and memory database as solutions to complex process in distribution system.Moreover,it also intro-duces the AEBrain(AliExpress AI application platform),ODPS(Open Data Process-ing Service)and other intelligent application platforms suitable for big data scenarios.The OneLink system successfully realizes the paid traffic intervention without communication between the advertising platform and promoters.The system has been put into operation,and has supported several big promotions' needs and daily needs.Practice showed that the system significantly improved the flow conversion data in the process of cooperating with the operation.