Mastering Apache Flink

Definitive guide to lightning fast data processing for distributed systems with Apache FlinkAbout This Book* Build your experitse in processing realtime data with Apache Flink and its ecosystem* Gain insights into the working of all components of Apache Flink such as FlinkML, Gelly, and Table APIFilled with real world use cases,* Your guide to take advantage of Apache Flink for solving real world problemsWho This Book Is ForBig data developers who are looking to process batch and real-time data on distributed systems. Basic knowledge of Hadoop and big data is assumed. Reasonable knowledge of Java or Scala is expected.What You Will Learn* Learn how to build end to end real time analytics projects* Integrate with existing big data stack and utilize existing infrastructure.* Build predictive analytics applications using FlinkML* Use graph library to perform graph querying and search.In DetailWith the advent of massive computer systems, organizations in different domains generate large amounts of data at a realtime basis. The latest entrant to big data processing, Apache Flink, is designed to process continuous streams of data at a lightning fast pace.This book will be your definitive guide to batch and stream data processing with Apache Flink. The book begins with introducing the Apache Flink ecosystem, setting it up and using the DataSet and DataStream API for processing batch and streaming datasets. Bringing the power of SQL to Flink, this book will then explore the Table API for querying and manipulating data. In the latter half of the book, readers will get to learn the remaining ecosystem of Apache Flink to achieve complex tasks such as event processing, machine learning, and graph processing. The final part of the book would consist of topics such as scaling Flink solutions, performance optimization and integrating Flink with other tools such as ElasticSearch.Whether you want to dive deeper into Apache Flink, or want to investigate how to get more out of this powerful technology, you'll find everything inside Cover Copyright Credits About the Author About the Reviewers www.PacktPub.com Customer Feedback Table of Contents Preface Chapter 1: Introduction to Apache Flink History Architecture Distributed execution Job Manager Actor system Scheduler Check pointing Task manager Job client Features High performance Exactly-once stateful computation Flexible streaming windows Fault tolerance Memory management Optimizer Stream and batch in one platform Libraries Event time semantics Quick start setup Pre-requisite Installing on Windows Installing on Linux Cluster setup SSH configurations Java installation Flink installation Configurations Starting daemons Adding additional Job/Task Managers Stopping daemons and cluster Running sample application Summary Chapter 2: Data Processing Using the DataStream API Execution environment Data sources Socket-based File-based Transformations Map FlatMap Filter KeyBy Reduce Fold Aggregations Window Global windows Tumbling windows Sliding windows Session windows WindowAll Union Window join Split Select Project Physical partitioning Custom partitioning Random partitioning Rebalancing partitioning Rescaling Broadcasting Data sinks Event time and watermarks Event time Processing time Ingestion time Connectors Kafka connector Twitter connector RabbitMQ connector ElasticSearch connector Embedded node mode Transport client mode Cassandra connector Use case – sensor data analytics Summary Chapter 3: Data Processing Using the Batch Processing API Data sources File-based Collection-based Generic sources Compressed files Transformations Map Flat map Filter Project Reduce on grouped datasets Reduce on grouped datasets by field position key Group combine Aggregate on a grouped tuple dataset MinBy on a grouped tuple dataset MaxBy on a grouped tuple dataset Reduce on full dataset Group reduce on a full dataset Aggregate on a full tuple dataset MinBy on a full tuple dataset MaxBy on a full tuple dataset Distinct Join Cross Union Rebalance Hash partition Range partition Sort partition First-n Broadcast variables Data sinks Connectors Filesystems HDFS Amazon S3 Alluxio Avro Microsoft Azure storage MongoDB Iterations Iterator operator Delta iterator Use case – Athletes data insights using Flink batch API Summary Chapter 4: Data Processing Using the Table API Registering tables Registering a dataset Registering a datastream Registering a table Registering external table sources CSV table source Kafka JSON table source Accessing the registered table Operators The select operator The where operator The filter operator The as operator The groupBy operator The join operator The leftOuterJoin operator The rightOuterJoin operator The fullOuterJoin operator The union operator The unionAll operator The intersect operator The intersectAll operator The minus operator The minusAll operator The distinct operator The orderBy operator The limit operator Data types SQL SQL on datastream Supported SQL syntax Scalar functions Scalar functions in the table API Scala functions in SQL Use case – Athletes data insights using Flink Table API Summary Chapter 5: Complex Event Processing What is complex event processing? Flink CEP Event streams Pattern API Begin Filter Subtype OR Continuity Strict continuity Non-strict continuity Within Detecting patterns Selecting from patterns Select flatSelect Handling timed-out partial patterns Use case – complex event processing on a temperature sensor Summary Chapter 6: Machine Learning Using FlinkML What is machine learning? Supervised learning Regression Classification Unsupervised learning Clustering Association Semi-supervised learning FlinkML Supported algorithms Supervised learning Support Vector Machine Multiple Linear Regression Optimization framework Recommendations Alternating Least Squares Unsupervised learning k Nearest Neighbour join Utilities Data pre processing and pipelines Polynomial features Standard scaler MinMax scaler Summary Chapter 7: Flink Graph API - Gelly What is a graph? Flink graph API – Gelly Graph representation Graph nodes Graph edges Graph creation From dataset of edges and vertices From dataset of tuples representing edges From CSV files From collection lists Graph properties Graph transformations Map Translate Filter Join Reverse Undirected Union Intersect Graph mutations Neighborhood methods Graph validation Iterative graph processing Vertex-Centric iterations Scatter-Gather iterations Gather-Sum-Apply iterations Use case – Airport Travel Optimization Summary Chapter 8: Distributed Data Processing with Flink and Hadoop Quick overview of Hadoop HDFS YARN Flink on YARN Configurations Starting a Flink YARN session Submitting a job to Flink Stopping Flink YARN session Running a single Flink job on YARN Recovery behavior for Flink on YARN Working details Summary Chapter 9: Deploying Flink on Cloud Flink on Google Cloud Installing Google Cloud SDK Installing BDUtil Launching a Flink cluster Executing a sample job Shutting down the cluster Flink on AWS Launching an EMR cluster Installing Flink on EMR Executing Flink on EMR-YARN Starting a Flink YARN session Executing Flink job on YARN session Shutting down the cluster Flink on EMR 5.3+ Using S3 in Flink applications Summary Chapter 10: Best Practices Logging best practices Configuring Log4j Configuring Logback Logging in applications Using ParameterTool From system properties From command line arguments From .properties file Naming large TupleX types Registering a custom serializer Metrics Registering metrics Counters Gauges Histograms Meters Reporters Monitoring REST API Config API Overview API Overview of the jobs Details of a specific job User defined job configuration Back pressure monitoring Summary Index

Discover the definitive guide to crafting lightning-fast data processing for distributed systems with Apache Flink

About This Book

• Build your expertize in processing real-time data with Apache Flink and its ecosystem
• Gain insights into the working of all components of Apache Flink such as FlinkML, Gelly, and Table API filled with real world use cases
• Exploit Apache Flink's capabilities like distributed data streaming, in-memory processing, pipelining and iteration operators to improve performance.
• Solve real world big-data problems with real time in-memory and disk-based processing capabilities of Apache Flink.

Who This Book Is For

Big data developers who are looking to process batch and real-time data on distributed systems. Basic knowledge of Hadoop and big data is assumed. Reasonable knowledge of Java or Scala is expected.

What You Will Learn

• Learn how to build end to end real time analytics projects
• Integrate with existing big data stack and utilize existing infrastructure
• Build predictive analytics applications using FlinkML
• Use graph library to perform graph querying and search.
• Understand Flink's - "Streaming First" architecture to implementing real streaming applications
• Learn Flink Logging and Monitoring best practices in order to efficiently design your data pipelines
• Explore the detailed processes to deploy Flink cluster on Amazon Web Services(AWS) and Google Cloud Platform (GCP).

In Detail

With the advent of massive computer systems, organizations in different domains generate large amounts of data on a real-time basis. The latest entrant to big data processing, Apache Flink, is designed to process continuous streams of data at a lightning fast pace.

This book will be your definitive guide to batch and stream data processing with Apache Flink. The book begins with introducing the Apache Flink ecosystem, setting it up and using the DataSet and DataStream API for processing batch and streaming datasets. Bringing the power of SQL to Flink, this book will then explore the Table API for querying and manipulating data. In the latter half of the book, readers will get to learn the remaining ecosystem of Apache Flink to achieve complex tasks such as event processing, machine learning, and graph processing. The final part of the book would consist of topics such as scaling Flink solutions, performance optimization and integrating Flink with other tools such as ElasticSearch.

Whether you want to dive deeper into Apache Flink, or want to investigate how to get more out of this powerful technology, you'll find everything you need inside.

Style and approach

This book is a comprehensive guide that covers advanced features of the Apache Flink, and communicates them with a practical understanding of the underlying concepts for how, when, and why to use them.

Annotation Discover the definitive guide to crafting lightning-fast data processing for distributed systems with Apache Flink About This Book - Build your expertize in processing real-time data with Apache Flink and its ecosystem - Gain insights into the working of all components of Apache Flink such as FlinkML, Gelly, and Table API filled with real world use cases - Exploit Apache Flink's capabilities like distributed data streaming, in-memory processing, pipelining and iteration operators to improve performance. - Solve real world big-data problems with real time in-memory and disk-based processing capabilities of Apache Flink. Who This Book Is For Big data developers who are looking to process batch and real-time data on distributed systems. Basic knowledge of Hadoop and big data is assumed. Reasonable knowledge of Java or Scala is expected. What You Will Learn - Learn how to build end to end real time analytics projects - Integrate with existing big data stack and utilize existing infrastructure - Build predictive analytics applications using FlinkML - Use graph library to perform graph querying and search. - Understand Flink's - "Streaming First" architecture to implementing real streaming applications - Learn Flink Logging and Monitoring best practices in order to efficiently design your data pipelines - Explore the detailed processes to deploy Flink cluster on Amazon Web Services(AWS) and Google Cloud Platform (GCP). In Detail With the advent of massive computer systems, organizations in different domains generate large amounts of data on a real-time basis. The latest entrant to big data processing, Apache Flink, is designed to process continuous streams of data at a lightning fast pace. This book will be your definitive guide to batch and stream data processing with Apache Flink. The book begins with introducing the Apache Flink ecosystem, setting it up and using the DataSet and DataStream API for processing batch and streaming datasets. Bringing the power of SQL to Flink, this book will then explore the Table API for querying and manipulating data. In the latter half of the book, readers will get to learn the remaining ecosystem of Apache Flink to achieve complex tasks such as event processing, machine learning, and graph processing. The final part of the book would consist of topics such as scaling Flink solutions, performance optimization and integrating Flink with other tools such as ElasticSearch. Whether you want to dive deeper into Apache Flink, or want to investigate how to get more out of this powerful technology, you'll find everything you need inside. Style and approach This book is a comprehensive guide that covers advanced features of the Apache Flink, and communicates them with a practical understanding of the underlying concepts for how, when, and why to use them About This BookBuild your expertize in processing real-time data with Apache Flink and its ecosystemGain insights into the working of all components of Apache Flink such as FlinkML, Gelly, and Table API filled with real world use casesExploit Apache Flink's capabilities like distributed data streaming, in-memory processing, pipelining and iteration operators to improve performance.Solve real world big-data problems with real time in-memory and disk-based processing capabilities of Apache Flink.Who This Book Is ForBig data developers who are looking to process batch and real-time data on distributed systems. Basic knowledge of Hadoop and big data is assumed. Reasonable knowledge of Java or Scala is expected.What You Will LearnLearn how to build real-time, end-to-end analytics projectsIntegrate with existing big data stacks and utilize existing infrastructuresBuild predictive analytics applications using FlinkMLUse the graph library to perform graph querying and searchesUnderstand Flink's -'Streaming First'architecture to implementing real streaming applicationsLearn Flink Logging and Monitoring best practices in order to efficiently design your data pipelinesExplore the detailed processes to deploy Flink cluster on Amazon Web Services(AWS) and Google Cloud Platform (GCP)In DetailWith the advent of massive computer systems, organizations in different domains generate large amounts of data on a real-time basis. The latest entrant to big data processing, Apache Flink, is designed to process continuous streams of data at a lightning fast pace.This book will be your definitive guide to batch and stream data processing with Apache Flink. The book begins with introducing the Apache Flink ecosystem, setting it up and using the DataSet and DataStream API for processing batch and streaming datasets. Bringing the power of SQL to Flink, this book will then explore the Table API for querying and manipulating data. In the latter half of the book, readers will get to learn the remaining ecosystem of Apache Flink to achieve complex tasks such as event processing, machine learning, and graph processing. The final part of the book would consist of topics such as scaling Flink solutions, performance optimization and integrating Flink with other tools such as ElasticSearch.Whether you want to dive deeper into Apache Flink, or want to investigate how to get more out of this powerful technology, you'll find everything you need inside.