# Amazon EMR

> The AWS service that runs Spark and Hadoop clusters for you.

**Field:** Data Engineering  
**Category:** Processing  
**Also known as:** AWS EMR, Amazon Elastic MapReduce, Elastic MapReduce, EMR

## Think of it like

It is like a rented workshop that comes with all the big tools set up.

## Definition

Amazon EMR is a cloud service from AWS that runs big data tools for you. It used to be called Amazon Elastic MapReduce. It sets up clusters for Apache Spark, Hadoop and similar tools. Data usually sits in Amazon S3, and clusters come and go around it. You pay for the machines while a cluster runs.

## Example

A team starts an EMR cluster each night. It runs its Spark jobs on data in S3, then shuts the cluster down by morning.

## Why it matters

Building a Spark cluster by hand is slow, fiddly work. EMR does it for you, so a team can focus on the jobs.

## Common mistake

Leaving a cluster running after the jobs are done. You pay for every machine for every hour it is on.

## Three principles

- **Managed clusters.** AWS sets up the machines.
- **Open tools.** Spark, Hadoop and more.
- **Data in S3.** Clusters come and go around it.

## Related terms

- [Apache Spark](https://learn.benibix.com/term/data/apache-spark)
- [Apache Hadoop](https://learn.benibix.com/term/data/apache-hadoop)
- [Amazon S3](https://learn.benibix.com/term/data/amazon-s3)
- [AWS Glue](https://learn.benibix.com/term/data/aws-glue)
- [Databricks](https://learn.benibix.com/term/data/databricks)
- [Amazon Web Services (AWS)](https://learn.benibix.com/term/infrastructure/amazon-web-services)

## Told apart from

- [Amazon EMR vs AWS Glue](https://learn.benibix.com/compare/amazon-emr-vs-aws-glue): People think Amazon EMR and AWS Glue are the same Spark service, and they are not. EMR builds clusters you can tune and run many tools on. You pay for the machines while they run. Glue is serverless. You hand it an ETL job, and AWS runs the Spark for you. Pick EMR for control and Glue for less work.

![A highlighted driver at the top runs your code and splits it into tasks. Arrows marked tasks lead down to three executors. Each executor holds its own slice of the big table and a highlighted copy of the small table. A note says that in a broadcast join each executor gets the small table and joins locally, so the big table never moves between machines.](https://learn.benibix.com/diagrams/dat3-spark-cluster.svg)

---

Source: https://learn.benibix.com/term/data/amazon-emr
Author: Behnia Zarei (a developer in Vienna, Austria, who builds and teaches data, AI and business intelligence at BENIBIX)
Publisher: BENIBIX Learn (https://learn.benibix.com/)
Last updated: 2026-11-19
Licence: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Quote freely with attribution.
