Research And Application Distributed Memory Database Based On Redis Cluster

As an open source of key-value model distributed memory database,redis cluster using distributed manner stored all user data in the node of memory,which greatly enhance the efficiency of data read and write operations,and this way meet user new requirements for traditional relational database.The paper based on the redis-3.0.3 version of source code design and implementation.Firstly we analyzed the principle of redis cluster,including redis databases,objects,data structures and also the design of distributed memory database.We also analyzed the master-slave synchronization,data persistence solution of aof and rdb.Secondly,we analyzed the problems of redis cluster and propose corresponding solutions.The content including dictionary rehash optimized,set of integers optimization,compression list chain updated optimization.And also including the design and implementation of part re-synchronization backlog buffer elastic expansion and complete re-synchronization rdb file incremental synchronization.We also proposed the solutions of hiredis regarding moved redirect and ask error handling,as well as implementing the cluster command-line tool redis-trib password integration.Then,we design test plan and use case data comparative the efficiency of redis cluster handle the same issues before and after optimization.The test plan including dictionary rehash testing,integers performance testing,compression list performance testing.And also including part re-synchronization copy backlog buffer elastic expansion testing,complete re-synchronization rdb file incremental synchronization testing,and the redis-trib tools password integration testing.Finally,based on the optimized redis-3.0.3 source code,build redis cluster tools redis-trib and optimized hiredis,we design and implementation heat url query system.And then we analysis the results of the system application,summarizes the existing problems for redis cluster practical application.