关键词: NoSQL, MongoDB, MapReduce, 聚合计算, 性能优化方法

Abstract: Abstract: In the big data application, based on NoSQL OLAP provide basic analysis capability. MongoDB is one of NoSQL, its built-in aggregation using basic MapReduce. On the one hand, database can improve the performance of aggregation by scale-clusters. On the other hand, database configuration method combined on application business can improve computing performance in the limited resources. This article is based from the practical application; an optimization method of MongoDB aggregation performance is proposed. The basic idea is to set to start from the shard key select and index set, choose the appropriate shard key and indexes by analysis shard feature. This method can make full use of existing resources to enhance the system aggregation performance. Through practical experiments related datasets, analysis and interpretation of experimental data to validate the method, it establishes a solid foundation for further study.

Key words: NoSQL(Not Only SQL), MongoDB, MapReduce, Aggregation, performance optimization