Course Description

Big Data on AWS

Course Overview

In this course, you’ll learn about cloud-based big data solutions like Amazon EMR, Amazon Redshift, Amazon Kinesis, and the rest of the AWS big data platform. Learn to use Amazon EMR to process data using the broad ecosystem of Hadoop tools like Hive and Hue, create big data environments, work with Amazon DynamoDB, Amazon Redshift, Amazon QuickSight, Amazon Athena and Amazon Kinesis, and design big data environments for security and cost-effectiveness.

  • Course level: Intermediate
  • Duration: 3 days

Activities

This course includes presentations, group exercises, and hands-on labs.

Course Objectives

In this course, you will learn to:

  • Use Apache Hadoop with Amazon EMR
  • Launch and configure an Amazon EMR cluster
  • Use common programming frameworks for Amazon EMR, including Hive, Pig, and Streaming
  • Use Hue to improve the ease-of-use of Amazon EMR
  • Use in-memory analytics with Spark on Amazon EMR
  • Understand how services like AWS Glue, Amazon Kinesis, Amazon Redshift, Amazon Athena, and Amazon QuickSight can be used with big data workloads

Intended Audience

This course is intended for:

  • Individuals responsible for designing and implementing big data solutions, namely Solutions Architects and SysOps Administrators
  • Data Scientists and Data Analysts interested in learning about big data solutions on AWS

Prerequisites

We recommend that attendees of this course have:

  • Basic familiarity with big data technologies, including Apache Hadoop, HDFS, and SQL/NoSQL querying
  • Completed Data Analytics Fundamentals free digital training or equivalent experience
  • Working knowledge of core AWS services and public cloud implementation
  • Completed the AWS Technical Essentials classroom training or have equivalent experience
  • Basic understanding of data warehousing, relational database systems, and database design

Get Started with Big Data on AWS

Contact us today to get started with AWS Classroom Training

Big Data on AWS Course Outline

Day 1

Module 1: Overview of Big Data

  • What is big data
  • The big data pipeline
  • Big data architectural principals

Module 2: Big Data Ingestion and Transfer

  • Overview: Data ingestion
  • Transferring data

Module 3: Big data streaming and Amazon Kinesis

  • Stream processing of big data
  • Amazon Kinesis
  • Amazon Kinesis Data Firehose
  • Amazon Kinesis Video Streams
  • Amazon Kinesis Data Analytics
  • Hands-on lab 1: Streaming and Processing Apache Server Logs Using Amazon Kinesis

Module 4: Big data storage solutions

  • AWS data storage options
  • Storage solutions concepts
  • Factors in choosing a data store

Module 5: Big data processing and analytics

  • Big data processing and analytics
  • Amazon Athena
  • Hands-on lab 2: Using Amazon Athena to Analyze Log Data
Day 2

Module 6: Apache Hadoop and Amazon EMR

  • Introduction to Amazon EMR and Apache Hadoop
  • Best practices for ingesting data
  • Amazon EMR
  • Amazon EMR architecture
  • Hands-on lab 3: Storing and Querying Data on Amazon DynamoDB

Module 7: Using Amazon EMR

  • Developing and running your application
  • Launching your cluster
  • Handling output from your completed jobs

Module 8: Hadoop programming frameworks

  • Hadoop frameworks
  • Other frameworks for use on Amazon EMR
  • Hands-on lab 4: Processing Server Logs with Hive on Amazon EMR

Module 9: Web interfaces on Amazon EMR

  • Hue on Amazon EMR
  • Monitoring your cluster
  • Hands-on lab 5: Running Pig Scripts in Hue on Amazon EMR

Module 10: Apache Spark on Amazon EMR

  • Apache Spark
  • Using Spark
  • Hands-on lab 6: Processing NY Taxi Data Using Apache Spark
Day 3

Module 11: Using AWS Glue to automate ETL workloads

  • What is AWS Glue?
  • AWS Glue: Job orchestration

Module 12: Amazon Redshift and big data

  • Data warehouses vs. traditional databases
  • Amazon Redshift
  • Amazon Redshift architecture

Module 13: Securing your Amazon deployments

  • Securing your Amazon deployments
  • Amazon EMR security overview
  • AWS Identity and Access Management (IAM) overview
  • Securing data
  • Amazon Kinesis security overview
  • Amazon DynamoDB security overview
  • Amazon Redshift security overview

Module 14: Managing big data costs

  • Total cost considerations for Amazon EMR
  • Amazon EC2 pricing models
  • Amazon Kinesis pricing models
  • Cost considerations for Amazon DynamoDB
  • Cost considerations and pricing models for Amazon Redshift
  • Optimizing cost with AWS

Module 15: Visualizing and orchestrating big data

  • Visualizing big data
  • Amazon QuickSight
  • Orchestrating a big data workflow
  • Hands-on lab 7: Using TIBCO Spotfire to visualize data

Module 16: Big data design patterns

  • Common architectures

Module 17: Course wrap-up

  • What’s next?
Contact Us to Learn More About

Big Data on AWS