EC2 Placement Groups

Nov 5, 2022

# EC2 Placement Group

Generally, instances in this placement group will use the same rack, and sometime even the same host.
Maximizes performance, all members have direct connections to each other.
10 Gbps per stream compared to the 5 Gbps which is achievable normally
Lowest latency and max PPS possible in AWS
Offers little to no resilience
Can’t span AZs - ONE AZ ONLY - locked when launching first instance
Can span VPC peers - but impacts performance
Requires a supported instance type
Use the same type of instance (not mandatory)
Launch at the same time (not mandatory … very recommended)
Use case: Performance, fast speeds, low latency

Keep instances separated
Ensure maximum amount of availability and resiliency for an application
Located on isolated, distinct racks per AZ
- Max of 7 instance per AZ, due to distinct rack placement
Provides infrastructure isolation
Each instance runs from a different rack
Each rack has its own network and power source
Not supported for Dedicated Instances or Hosts
Use Case: Small number of critical instances that need to be kept separated from each other

Groups of instances, spread apart
Similar to spread placement groups, but for situations where you have more than 7 instances per AZ. There is no instance limit per AZ.
Divided into partitions, max 7 partitions per AZ
Each partition has its own racks, no sharing between partitions
Designed for huge scale parallel processing systems
You can control which instances are placed in which partitions - Topology Aware
Instances can be placed in a specific partition or auto placed
HDFS, HBase, and Cassandra