The Design And Application Of System Architecture That Jackrabbit Encapsulating MongoDB As Basic Storage

With the popularity of the Internet and the rapid development of computer technology,this situation results in the accumulation of a large number of network data.These huge amounts of data information includes structured data and unstructured data.It has become a problem which cannot be ignored to store and manage these heteroid mass data.At present,the CMS manufacturers can use Java content repository standard interfaces in a neutral way to access their own repository.However,the underlying storage medium of Java content repository standard specification doesn’t support NoSQL database such as MongoDB.Under such background environment,this paper studies and analyses the internal components of Apache Jackrabbit,basic storage of Jackrabbit supporting,the process of encapsulating basic storage and how to build a high availability of distributed cluster with MongoDB.In the design that Jackrabbit encapsulates MongoDB database,node and property are stored with PersistenceManagement as well as some binary file.The Large binary files are stored with DataStore. Encapsulation of PersistenceManagement takes advantage of BSON document of MongoDB. The GridFS storage mechanism of MongoDB can be using to resolve the problem of encapsulation of the DataStore.The boundary of binary file size can be determined by experimental test,so that the system can get the best storage performance.Jackrabbit system architecture encapsulating MongoDB improves the performance of the mass data storage with JSR-170.The building of distributed high availability cluster adopts the combination of Sharding and Replica Set.This cluster improves high availability,scalability and flexibility of system.This distributed cluster includes three mongos router servers,three configuration servers and three shards.Every shard is a replica set.There is one primary node,one secondary node and one arbitration node in every replica set.If the primary is broken down,the arbitration node will choose a secondary as primary according to the Bully algorithm.The operation of writing data only act on the primary,and data can be read from secondary,as a result that this implements to read and write data concurrently.This distributed cluster can increase the storage capacity of the system by adding shards and improve the reliability of the system by adding secondary in replica set.This article mainly includes the design and implementation of encapsulating and the application in university website content management system.The first chapter mainly summarizes the cause and development background of researching.The second introduces the core technology of this paper.The third mainly illustrates the requirement analysis of system architecture.Fourth and fifth chapter is the key chapter of this paper,which describes in detail the encapsulation process of Jackrabbit for MongoDB,the use of core interface in encapsulation and the application of jackrabbit system architecture in university website content management system.The last chapter is a summary of this paper and the prospect of the next step work.