Servisphere
Amazon EMR icon (AWS Architecture Icons)

Amazon EMR

Spark/Hadoop big data clusters

Runs big data frameworks such as Apache Spark, Hive, Presto, and Flink on EC2 (or EKS) clusters that AWS sets up for you; data usually lives in S3.

  • Cost$$ Medium
  • Difficulty●●●●●
  • Costs money when idle?Yes
  • Managed by AWS
  • In examsSAA

Scope: Regional (cluster in one AZ)

Open EMR on the map →
Think of it asLike renting an entire factory floor for mass production, then handing it back when the job is done.

When to use it

✓ Use it when

Processing big data with Spark/Hadoop when you need to customize the cluster.

✗ Avoid it when

Don't want to manage clusters → EMR Serverless / Glue.

Common uses

  • ETL on terabytes of data with Spark
  • Move on-prem Hadoop clusters to AWS
  • Distributed ML training on big data

What does it cost?

  • Pricing model: Charged per hour of instance runtime (EC2 fee plus an EMR surcharge per instance).
⚠️ Cost traps

Use Spot for task nodes and auto-terminate clusters when jobs finish to avoid paying for idle clusters.

Reference prices for us-east-1, updated October 14, 2026; they may have changed. Always check the AWS pricing page.

Related services

SAA-C03 study guide

EMR is covered in these study modules:

Open EMR on the map →