Redshift Observatory

Cross-Region Benchmarks

The benchmark suite is run bi-weekly, and checks almost all regions (not Gov, not China, not regions where RS cluster do not in fact start).

The benchmarks are for disk(r), disk(rw), network, processor.

disk(r) : Sequential read 512mb.
disk(rw) : Sequential read and then sequential write 512mb.
network : Sequential read 512mb and send to another node.
processor : Generate 2^23 single column rows in memory.

The benchmarks are described in detail in the white paper Cross-Region Benchmarks.

Usually, only the three small node types are benchmarked, due to cost.

Clusters are made of nodes, but node performance varies greatly; you can't really test a cluster, because all you're really doing is testing the speed of the slowest node. All tests then are issued on a single slice of a single node (except the network test, which uses two nodes, one to send, one to receive).

All benchmarks are for execution time only and are taken from the Redshift system tables.

All benchmarks are repeated five times, with the slowest and fastest durations discarded.

Benchmarking Changes
date	change
2025-02-01	As of early 2025 disk encryption has been made mandatory, and so from 2025-02-01 and onward, is enabled.

	dc2.large	dc2.8xlarge	ra3.large	ra3.xlplus	ra3.4xlarge	ra3.16xlarge
af-south-1
ap-east-1
ap-northeast-1
ap-northeast-2
ap-northeast-3
ap-south-1
ap-south-2
ap-southeast-1
ap-southeast-2
ap-southeast-3
ap-southeast-4
ca-central-1
cn-north-1
cn-northwest-1
eu-central-1
eu-central-2
eu-north-1
eu-south-1
eu-south-2
eu-west-1
eu-west-2
eu-west-3
me-central-1
me-south-1
sa-east-1
us-east-1
us-east-2
us-gov-east-1
us-gov-secret-1
us-gov-topsecret-1
us-gov-topsecret-2
us-gov-west-1
us-west-1
us-west-2