The Research Of High Availability And Performance Optimization Of Distributed MongoDB Clusters

Along with the rapid development of network technology, the data volume and traffic are rising rapidly, the amount of data on internet has changed from MB,GB to TB.The scale of various kinds applications and websites is very large, concurrency is huge,and the dynamic resource of the websites is liberal, the data format is complex and irregular.The design of the database is also becoming more complex,the joint query of the databases is also increasing,faced with these demands on the function and performance,traditional databases become stretched in some scenarios.MongoDB uses the Nosql pattern,its the document model is free.For the internet application of large amount of data,high concurrency and eventual consistency,MongoDB can have very good performance.In the case of the rising of data volume,it can achieve rapid and convenient expansion,the performance is also very good.Built-in horizontal expansion mechanism provides a capability of data processing that amount from 1 million to 1 billion,Mongo DB completely meet the demand of the data storage for web2.0 and mobile internet,out-of-the-box feature also greatly reduces the cost of operation and maintenance of small and medium websites.In view of the above background and application requirements, based on the outstanding characteristics of Mongo DB,in this paper, we use the Docker virtualization solutions, design a distributed cluster system of high availability, realized the high availability and high performance data scheme, solve the problems such as high cost of bank log information storage, complex relationship and data island,and pass the relevant function and performance test,achieve the standard of application.In the process of implementation, through the analysis of the application requirements, the paper designs a distributed architecture with disaster recovery, high performance and high scalability.In the specific implementation process,the system use Docker to build distributed nodes,after the building,the routing node deploy these nodes with different functions.Finally, a distributed database cluster with high availability and scalability will be formed,the data of the project is analyzed at the same time.According to the requirement of the business, the data storage scheme is designed for the cluster.In the application of database, optimization has always been the focus of research.Based on the above data solution, this paper studies the source code in-depth,optimizes the whole cluster from the aspects of software and hardware in the case of understanding its mechanism, and tests its performance to confirm whether can achieve the corresponding indicators.Distributed database is different from single point database,the data operation of cluster involves the balance and integration of data between multiple nodes.Through in-depth studying of Mongo DB MapReduce parallel programming and channel aggregation programming,this paper discusses the main technical points of the two schemes,designs process according to the specific application requirements,analyzes the application scenario of two schemes, puts forward a scheme of distributed query that suitable for wide scene.