主要特点
|
Best of BigTable and Dynamo
|
Retains some friendly properties of SQL
|
DB consistency,ease of use
|
Blazing fast
|
Fault tolerence
|
Billions of rows ,millions of columns
|
适用场景
|
熟悉Bigtable,数据读取量大于写,组件须由Java开发
|
动态查询、数据经常填满磁盘、另外用户喜欢自定义索引而非Mapreduce
|
大量数据,但更新量小,预定义查询
|
数据快速变化,但数据库大小可预见(适合内存存取数据)
|
数据快速膨胀和数据复杂性。容错性及可扩展性需要更多服务器支持
|
随机数据,随机读写海量数据
|
用例
|
Banking,Financial Industry
|
For all things that you would do with MySQL or PostgreSQL,but having predefined columns really holds you back
|
CRM、CMS systems,Master-master replication is a especially interesting feature,allowing easy multi-site
|
Stock prices,Analytics,Real-time data collection,Real-time Communication
|
Point-of-sales data collection,Factory control systems, Places where even seconds of downtime hurt
|
Facebook Messaging Database
|