Book Image

Learning Big Data with Amazon Elastic MapReduce

By : Amarkant Singh, Vijay Rayapati
Book Image

Learning Big Data with Amazon Elastic MapReduce

By: Amarkant Singh, Vijay Rayapati

Overview of this book

<p>Amazon Elastic MapReduce is a web service used to process and store vast amount of data, and it is one of the largest Hadoop operators in the world. With the increase in the amount of data generated and collected by many businesses and the arrival of cost-effective cloud-based solutions for distributed computing, the feasibility to crunch large amounts of data to get deep insights within a short span of time has increased greatly.</p> <p>This book will get you started with AWS so that you can quickly create your own account and explore the services provided, many of which you might be delighted to use. This book covers the architectural details of the MapReduce framework, Apache Hadoop, various job models on EMR, how to manage clusters on EMR, and the command-line tools available with EMR. Each chapter builds on the knowledge of the previous one, leading to the final chapter where you will learn about solving a real-world use case using Apache Hadoop and EMR. This book will, therefore, get you up and running with major Big Data technologies quickly and efficiently.</p>
Table of Contents (18 chapters)
Learning Big Data with Amazon Elastic MapReduce
Credits
About the Authors
Acknowledgments
About the Reviewers
www.PacktPub.com
Preface
Index

Launching and monitoring an EMR cluster using CLI


Before launching an EMR cluster, you need to decide on the AWS region that will be used to launch the cluster and accordingly, you should have configured your credentials.json file. As discussed in our initial chapters, choosing a specific AWS region depends on factors such as your business location and latency requirements of connecting your existing data center or office with AWS using the virtual private network and so on for a secure data transfer.

Another important consideration is choosing the right instance type based on the analysis requirements. You would also need to consider an EMR cluster size depending on the size of data to be analyzed and stored in HDFS for processing. One m1.xlarge instance provides 1,680 GB of disk storage, so if you have an HDFS replication factor of 3, then you need at least three core nodes along with one master node for processing 1 TB of data. However, your cluster size also depends on the MapReduce job...