Skip to content

051 placement groups

Strategies

Cluster

  • single AZ (not single rack)
  • high per-flow throughput limit for TCP/IP

For

  • Apps that benefit from low network latency, high network throughput or both
  • when the majority of the traffic is between the instances in group

Limit:

  • using the same instance type for all instances is recommended.

Partition

spread across different partition

  • each partition within a placement group has its own set of racks

For:

  • large distributed and replicated workloads, such as HDFS, HBase, and Cassandra, across distinct racks

Limit:

  • maximum of seven partitions per Availability Zone

Spread

spread across different hardware (rack)

  • reduces the risk of simultaneous failures that might occur when instances share the same equipment

For:

  • mix instance types
  • launch instances over time

Limit:

  • up to 7 instances per AZ (no multiple region)