Researches On Redis Cache Cluster And Its Application In 3M Scheme

With the rapid development of cloud computing and Internet technology,more and more Web systems adopt SaaS platform,which uses Multi-Tenant technology that can provide services to different tenants.In order to let tenants better to customize different services based on their own requirements,Multi-Application Multi-Tenant Multi-Factor scheme based data platform came into being.This kind of data platform not only brings convenience to users,but also brings challenges to efficient access to data.With the increasing quantities of tenants and data,the system cannot respond quickly to reading and writing requests of users,especially when several tenants visit system at the same time,system response will become longer.In this thesis,a Multi-Application Multi-Tenant Multi-Factor scheme based data platform is taken as research background.In order to achieve high-speed data access and meet real-time requirements of tenants,Redis cache database based universal distributed caching system is investigated and constructed,since it is suitable for 3M data platform,while the he load balancing algorithm is optimized.This thesis discusses the following points:(1)It analyzes the existing disadvantages of data storage strategy of 3M data platform,compared with the properties of common cache database,and chooses Redis as the cache container.Meanwhile,it analyzes the limitations of Redis cluster in practical application and points out that using distributed service component Zookeeper to create high-available distributed caching system.(2)It analyzes the advantages and disadvantages of common used load algorithms,chooses the consistent hashing algorithm as the basic algorithm of load balance in the application scene of 3M data platform,and optimizes it based on the characteristics of the efficiency of accessing and load balancing.(3)According to the characteristic of data structure in 3M date platform,it designs and achieves the distributed caching system which is able to support the 3M scheme,including register management and monitor Redis node module according to distributed coordination component Zookeeper,and the connection to Redis in application by using Redis client technology Jedis to read and write.(4)To decouple the cache service and business logic,and reduce the workload of system development,it also designs and realizes the API of handling reading and writing requests which is able to support the 3M scheme.(5)It tests and verifies the function of the distributed caching system which is able to support the 3M scheme,and analyzes the performance of the improved algorithm.Finally,applies this caching system in SaaS-based Web systems and analyzes its performance.