Why Placement Groups?

By default, AWS decides where to physically place your EC2 instances. Placement Groups let you control that decision based on your needs — whether you want maximum speed, maximum safety, or large-scale distributed setups.

Types of Placement Groups

1. Cluster

All instances are packed together on the same hardware in a single Availability Zone.

• Pro: Ultra-low latency and high network speed between instances (up to 10 Gbps)
• Con: If the underlying hardware fails, all instances go down together
• Use for: HPC (High Performance Computing), ML training, big data processing jobs
• Trade-off: Speed over safety

2. Spread

Each instance is placed on completely separate hardware — different rack, different power source, different network.

• Pro: Maximum fault isolation — if one rack fails, only that one instance is affected
• Con: Maximum 7 instances per Availability Zone (AWS needs to reserve a unique rack per instance)
• Use for: Critical applications that cannot afford simultaneous failures — RDS, important app servers
• Trade-off: Safety over scale

3. Partition

Instances are divided into groups called partitions. Each partition runs on separate racks, but within a partition you can have many instances.