Get the maximum Amazon EBS optimized performance - Amazon Elastic Compute Cloud

Get the maximum Amazon EBS optimized performance

An instance's EBS performance is bounded by the instance type's performance limits, or the aggregated performance of its attached volumes, whichever is smaller. To achieve maximum EBS performance, an instance must have attached volumes that provide a combined performance equal to or greater than the maximum instance performance. For example, to achieve 80,000 IOPS for r6i.16xlarge, the instance must have at least 5 gp2 volumes provisioned with 16,000 IOPS each (5 volumes x 16,000 IOPS = 80,000 IOPS), or it can have 1 gp3 volume provisioned with 80,000 IOPS. We recommend that you choose an instance type that provides more dedicated Amazon EBS throughput than your application needs; otherwise, the connection between Amazon EBS and Amazon EC2 can become a performance bottleneck.

Important

When using configurable bandwidth weightings, the EBS bandwidth limits for your instance might change. For instances with the VPC-1 weighting configuration, which increases networking bandwidth, you might experience lower than expected IOPS for EBS volumes due to reaching the EBS bandwidth limit before the IOPS limit. This is particularly noticeable with larger I/O sizes. Always test your specific workload to ensure it meets your performance requirements with your selected bandwidth weighting. For more information, see EC2 instance bandwidth weighting configuration.

You can use the EBSIOBalance% and EBSByteBalance% metrics to help you determine whether your instances are sized correctly. You can view these metrics in the CloudWatch console and set an alarm that is triggered based on a threshold you specify. These metrics are expressed as a percentage. Instances with a consistently low balance percentage are candidates to size up. Instances where the balance percentage never drops below 100% are candidates for downsizing. For more information, see Monitor your instances using CloudWatch.

The high memory instances are designed to run large in-memory databases, including production deployments of the SAP HANA in-memory database, in the cloud. To maximize EBS performance, use high memory instances with an even number of io1 or io2 volumes with identical provisioned performance. For example, for IOPS heavy workloads, use four io1 or io2 volumes with 40,000 provisioned IOPS to get the maximum 160,000 instance IOPS. Similarly, for throughput heavy workloads, use six io1 or io2 volumes with 48,000 provisioned IOPS to get the maximum 4,750 MB/s throughput. For additional recommendations, see Storage Configuration for SAP HANA.

Considerations

  • G4dn, I3en, Inf1, M5a, M5ad, R5a, R5ad, T3, T3a, and Z1d instances launched after February 26, 2020 provide the maximum EBS optimized performance. To get the maximum performance from an instance launched before February 26, 2020, stop and start it.

  • C5, C5d, C5n, M5, M5d, M5n, M5dn, R5, R5d, R5n, R5dn, and P3dn instances launched after December 3, 2019 provide the maximum EBS optimized performance. To get the maximum performance from an instance launched before December 3, 2019, stop and start it.

  • u-6tb1.metal, u-9tb1.metal, and u-12tb1.metal instances launched after March 12, 2020 provide the maximum EBS optimized performance. Instances of these types launched before March 12, 2020 might provide lower performance. To get the maximum performance from an instance launched before March 12, 2020, contact your account team to upgrade the instance at no additional cost.