基于MongoDB的文件存储系统的设计与实现

Design and Implementation of File Storage System Based on Mongo DB

基于Mongo DB数据库,采用复制集群与分片集群结合的方式,建立分片复制集群(本文中为3分片,每个分片为3副本),集群包括路由节点、配置节点、主副本节点、仲裁节点.这样既兼顾了集群的性能,也保证了其高可用性.可有效解决原关系型数据库的不足,提高非结构化数据的查询效率,缓解关系数据库的读写负载压力和运维压力,而且当用户访问量剧增时,可以动态增加Mongo DB节点来扩展性能和提高负载能力.

Based on MongoDB, a kind of slice replication cluster can be created by the combination of replicationclusters and slice clusters (The slice replication cluster has three slices and every slice has three replicas). The creationof the slice replication cluster, which includes routing nodes, configuration nodes, primary replica nodes and arbitritionnodes, not only takes into accounts the performance of the cluster but also can ensure its high availability. It can effectivelysolve the problem of the insufficiency of the original relational database, improve the query efficiency of the unstructureddata, alleviate the read and write load pressure and operation pressure of the relational database, but also dynamicallyadd mongodb nodes to extend performance and increase load capacity when the number of access users increases.