Preface

        What I want to share with you this time is the big data stream processing engine , I had a brief understanding of it in the technical forum through the release of its new version , Always wanted to use , But looking back , No time 、 Second, there is no environment 、 Three no scenes . Finally, I can only get addicted in my mind , This time, there may be a rare opportunity to use the scene on the ground , Just for the time being, there is no urgent need on hand 、 The problem needs to be dealt with . Leaders give such an opportunity to learn and improve , You can enter the door first , Gradually improve in the later transformation process .
        in addition , Here's the thing , In fact, I'm too lazy . Obviously, I am CSDN Founding learning member of ,Flink The course also offers , And there are many , There are tutorials for high school and low school , Looked at the , The version is still new . In this tutorial, I read Mr. Li's basic article 100 lesson , It feels good . Thanks for that CSDN Pay in silence , A bridge built for our technicians .
        I sometimes say I can't bring others , In the eyes of others , It's not that people think so ？ In the future, we still have to get rid of our distractions , Tighten your heart , Create your own conditions ,good good study,day day up!

One 、Flink What is it? ？

It's said on the official website ：
       Apache Flink It is a distributed processing engine and framework for stateful computing on bounded and unbounded data streams .Flink Designed to run in all common cluster environments , Perform calculations at any size and memory level speed .

Two 、 Origin and development process

1. origin

        Time should be Hadoop When the fire started , So it wasn't taken seriously at that time .

2. framework

Big data processing 、 The process of machine learning engine

3. Mainstream engine comparison

Schools & Represents product comparison

Genre interpretation
Record ACK: feedback
Micro-batching: Micro batch
Transactional updates: Transaction update
Distributed snapshots: Distributed snapshots
Dataflow： The model proposed by Google , Model basis of stream processing

3、 ... and 、 Reasons for selection

1. Advantages

2. Represents product comparison

Comparison of operating environment

Ecological comparison

The organization and driving force behind

3.Flink

Application scenarios

Event driven

Event driven must be real-time . Let's do a calculation , Trigger an event 、 Write to downstream message queue ...

Data analysis

ETL The Conduit

Flink Of connectors

NIFI The project of the United States tourism and Aviation Administration , Use as a pipe . You can write data to other places through this pipeline , It doesn't have to be realized by yourself .
There are not enough pictures here , Sure github Search on flink、flinkx Look inside ,connectors More complete , The integration code inside is an example （ One of the learning methods recommended by Miss Li ）

Typical event analysis examples

summary

1、 Make a summary with Alibaba's use case

2、Flink A breakthrough in the future

3、 Who will be stronger in the future ？

I like Mr. Li's conclusion that who is strong and who is weak . We are human beings 、 Who is not a technical person ？