Design And Implementation Of User-Centric E-Commerce System In Distributed Environment

In recent years,with the rapid development of Internet technology,various forms of e-commerce system have emerged one after another.In practical applications,the scale of system users has generated a large amount of e-commerce data with business expansion.The traditional e-commerce system based on centralized architecture has been difficult to meet the requirements of mass e-commerce data storage and processing.In addition,in the face of large-scale e-commerce data,information overload becomes a normal state,and personalized recommendation technology can help users to extract valuable information from complicated data.However,the traditional personalized recommendation technology runs on a single machine.Due to the performance limitations of the single node,it is unable to process massive e-commerce data quickly and efficiently.A good data processing platform is needed to meet the processing requirements of massive e-commerce data.In order to solve the above problems,this paper designs a user-centric e-commerce system deployed in a distributed architecture.The system mainly includes a background management module and a front management module,and the front management module is separately divided into six sub-modules,including a front portal module,a search module,a member and a single sign-on module,a shopping cart module,an order module,and a personalized recommendation module.Each module is deployed separately on a separate server and invoked through an interface.In order to alleviate the database storage pressure,the background management module deploys a separate image server for storing product image data,and uses Nginx to provide http services and ftp to provide product image uploading services.In order to reduce the number of queries to the database,a separate Solr full-text search server is deployed in the search module,and the commodity search field is configured to implement the product search.A single sign-on interface is issued in the member and single sign-on module to ensure that users can log in to each module once.Personalized recommendation module In order to solve the problem of real-time efficiency of system resources and algorithms,the Hadoop platform is used for distributed storage and processing of e-commerce big data.By comparing K-means algorithm and CFSFDP algorithm,the improved CFSFDP algorithm is proposed.Combined with the actual e-commerce data set,the clustering algorithm and collaborative filtering algorithm are effectively combined to design a combined algorithm that satisfies the user's needs and is accurate and efficient.User's more efficient and accurate personalized recommendation function.Finally,the real data set is used to compare and verify the clustering algorithm.At the same time,the system is tested in various aspects of the simulation data.The system can well complete the functions determined by each module in the demand analysis,and the interface reserved by the system can be It is very convenient to expand the system business with considerable flexibility.