传统的关系数据库（ORDBMS）在面对并发量极高的Web2.0网站时已经显得捉襟见肘；而非关系型的数据库依靠其自身的优势却获得了快速的发展。其松散的数据结构、高效的数据处理能力、智能的数据分片功能等，都得到了越来越多的人的认可。首先介绍了MongoDB数据库的原理，通过分析源代码说明了 MongoDB 的自动分片（Auto-sharding）工作机制，并对其负载均衡算法提出改进，提出基于数据访问时间间隔（IODA)的负载均衡算法。%Abstra ct:Traditional database revealed a lot of problems when facing web2.0 sites with Large scale and high concurrency, while the Non -relational database has been developed rapidly because of its special features, such as loose data structure, effi-cient capacity of data processing, intelligent function of data partition and so on. At first, this paper intro-duces the principle of the MongoDB;then it analyzes the mechanism of auto-sharding in MongoDB by an-alyzing the source code; at last, an improved algo-rithm based on the Interval of Data Access is put forward. The improved balancing strategy can effi-ciently balance the data among shards.
展开▼
