| In recent years,with the popularity of the Internet,e-commerce has developed rapidly.Most e-commerce platforms have recognized the impact of reviews on sales and provide consumers with a feature-rich review system.The review system has some problems in the process of development,mainly because of the increase of users,the increase of the number of visits,the system's response time is too long,and even collapse.In order to solve these problems,it is necessary to design and implement a high concurrent review system in combination with specific business scenarios.This thesis mainly studies the database technology and clustering technology used in high concurrency optimization,and reviews related technologies such as distributed service framework Dubbo,distributed cache Redis,message queue Kafka and distributed search engine Elasticsearch.Based on the research of related technologies,this thesis designs and implements a highly concurrent e-commerce review system.The system adopts a hierarchical structure,which mainly includes a data storage layer,a business layer,and a load balancing layer.The data storage layer uses Mycat to implement sub-database partitioning and read-write separation.The database uses a dual-master and dual-slave structure,and uses the redis cluster mode to build redis clusters,and uses Elasticsearch to store background data.The business layer divides the system into a review module,a tag module,a question and answer module,and a background management module,and implements service using Dubbo.In order to decouple the review module from the tag module,a message queue Kafka is introduced.For high-concurrency requests,the business layer uses a multi-level caching scheme.For requests with higher concurrency requirements,the data is stored in the local cache,and the local cache is implemented using Guava Cache.The load balancing layer uses LVS and Nginx to achieve load balancing,and LVS uses Keepalived to ensure high availability.The e-commerce review system designed and implemented in this thesis has been put into operation at present.It has passed two rounds of promotion and can cope with high concurrent access,which has certain application value. |
