当前位置:网站首页>When tidb and Flink are combined: efficient and easy to use real-time data warehouse
When tidb and Flink are combined: efficient and easy to use real-time data warehouse
2020-11-07 20:15:00 【InfoQ】
With the rapid development of Internet , There will be more and more kinds of business , The volume of business data will grow , When it reaches a certain scale , The traditional data storage structure can not meet the needs of enterprises , Real time data warehouse becomes a necessary basic service . In terms of dimension Join For example , Data is stored in a business data source in the form of a normal form table , A lot of Join operation , Reduce performance . If it can be completed in the process of data cleaning and importing Join, Then there is no need to analyze again Join, To improve query performance .
Using real-time data warehouse , Enterprises can achieve real-time OLAP analysis 、 Real time data Kanban 、 Real time business monitoring 、 Real time data interface service, etc . But think of real-time data warehouse , Many people's first impression is that the architecture is complex , Difficult to operate and maintain . And thanks to the new version Flink Yes SQL Support for , as well as TiDB HTAP Characteristics of , We explored an efficient 、 Easy-to-use Flink+TiDB Real time data warehouse solution .
This article will first introduce the concept of real-time data warehouse , Then introduce Flink+TiDB The architecture and advantages of real-time data warehouse , Then we give some user scenarios that are already in use , Finally, it is given in docker-compose In the environment Demo, For readers to try .
The concept of real-time data warehouse
The concept of data warehouse is in 90 Age from Bill Inmon Put forward , It refers to a topic oriented 、 Integrated 、 Relatively stable 、 A collection of historical changes , Used to support management decisions . The data warehouse at that time collected data from data sources through message queues , By calculating daily or weekly for use in reports , Also known as offline data warehouse .
Link to the original text :【https://www.infoq.cn/article/IoD228mbbr7wylDEQKkh】. Without the permission of the author , Prohibited reproduced .
版权声明
本文为[InfoQ]所创,转载请带上原文链接,感谢
边栏推荐
猜你喜欢
Application and principle of handlermethodargumentresolver
技术总监7年自述——如何选择一家好公司
Exclusive interview with Yue Caibo
C语言Ⅰ博客作业03
一万四千字分布式事务原理解析,全部掌握你还怕面试被问?
9. Routingmesh service communication between clusters
Knowledge competition of garbage classification
Didi's distributed ID generator (tinyid), easy to use
PHP安全:变量的前世今生
Rech8.0 learning days 12 rh134
随机推荐
Exception calling 'downloadstring' with '1' arguments: 'operation timed out'
Knowledge competition of garbage classification
ajax 载入html后不能执行其中的js解决方法
On the coverage technology and best practice of go code
如何以计算机的方式去思考
chrome浏览器跨域Cookie的SameSite问题导致访问iframe内嵌页面异常
Big data algorithm - bloon filter
把 4个消息队列都拉到一个群里后,他们吵起来了
【涂鸦物联网足迹】物联网主流通信方式
AC86U kx上网
Git代码提交操作,以及git push提示failed to push some refs'XXX'
On hiz buffer
30岁后,你还剩下什么?
graph generation model
CPU瞒着内存竟干出这种事
模型预测准确率高达94%!利用机器学习完美解决2000亿美元库存难题
HMS core push service helps e-commerce app to carry out refined operation
Web API系列(三)统一异常处理
在pandas中使用pipe()提升代码可读性
一万四千字分布式事务原理解析,全部掌握你还怕面试被问?