Skip to content

EC2 Placement Groups

Links: 101 AWS SAA Index


  • It helps us control how our EC2 instances are going to be placed within the AWS infrastructure.
  • We don’t get access to the hardware but we let AWS know how we want our EC2 machines to be placed.
  • There is no charge for creating a placement group

Cluster

  • Low latency and high throughput. Like HPC.
  • Placed in the same rack in the same AZ.
  • All instances are on the same hardware.

Spread

  • Minimise failure risk.
  • Limited 7 instances per AZ per placement group. So if you want 15 instances in spread placement group then you will need 3 AZs. Major CON.
  • Instances are on different hardware in different AZ.
  • Critical applications.

Partition

  • We have instances spread across partitions in multiple AZs.
  • We can have upto 7 partitions per AZ. Each partition can have many EC2 instances.
  • Each partition represents a hardware rack.
  • By having many partitions you are making sure that your EC2 instances are distributed across many hardware racks. Each partition is isolated from failure since they are on a different rack
  • We can have 100s of EC2 instances with this setup.
  • We can get partition information using the EC2 metadata service.
  • Use cases would be big data applications like Cassandra, HBase, Kafka
If you receive a capacity error when launching an instance in a placement group that already has running instances, stop and start all of the instances in the placement group, and try the launch again.
If you have an application that needs a large number of instances and minimise hardware failure then go for partition over spread.

Last updated: 2022-05-05