Storage - AWS Outposts High Availability Design and Architecture Considerations

This document is in the process of being updated. In the interim, some of the content might not be accurate.


The AWS Outposts rack service provides three storage types:

Instance storage is provided on supported servers (C5d, M5d, R5d, G4dn, and I3en). Just like in the Region, the data in an instance store persists only for the (running) lifetime of the instance.

Outposts EBS volumes and S3 on Outposts object storage are provided as part of the AWS Outposts rack managed services. Customers are responsible for capacity management of the Outpost storage pools. Customers specify their storage requirements for EBS and S3 storage when ordering an Outpost. AWS configures the Outpost with the number of storage servers required to provide the requested storage capacity. AWS is responsible for the availability of the EBS and S3 on Outposts storage services. Sufficient storage servers are provisioned to provide highly available storage services to the Outpost. Loss of a single storage server should not disrupt the services nor result in data loss.

You can use the AWS Management Console and CloudWatch metrics to monitor Outpost EBS and S3 on Outposts capacity utilization.

Data protection

For EBS Volumes: AWS Outposts rack supports EBS volume snapshots to provide a simple and secure data protection mechanism to protect your block storage data. Snapshots are point-in-time incremental backups of your EBS volumes. By default, snapshots of Amazon EBS volumes on your Outpost are stored on Amazon S3 in the Region. If your Outposts have been configured with S3 on Outposts capacity, you can use EBS Local Snapshots on Outposts to store snapshots locally on your Outpost using S3 on Outposts storage.

For S3 on Outposts buckets (data residency use cases):

  • You can use S3 Versioning on Outposts, to save all changes, and history of objects. When enabled, S3 Versioning saves multiple distinct copies of an object in the same bucket. You can use S3 Versioning to preserve, retrieve, and restore every version of every object stored in your Outposts buckets. S3 Versioning helps you recover from unintended user actions and application failures.

  • You can use S3 Replication on Outposts, to create and configure replication rules to automatically replicate your S3 objects to another Outpost, or to another bucket on the same Outpost. During replication, S3 on Outposts objects are sent over the customer's local gateway (LGW), and objects do not travel back to the AWS Region. S3 Replication on Outposts provides an easy and flexible way to automatically replicate data within a specific data perimeter to address data redundancy and compliance requirements.

    S3 Replication on Outposts also provides detailed metrics and notifications to monitor the status of your object replication. You can monitor replication progress by tracking bytes pending, operations pending, and replication latency between your source and destination Outposts buckets using Amazon CloudWatch. You can also set up Amazon EventBridge rules to receive replication failure events to quickly diagnose and correct configuration issues.

    Reference architecture diagram showing using Amazon EventBridge and Amazon CloudWatch

For S3 on Outposts buckets (non-data residency use cases) to AWS Regions: You can use AWS DataSync to automate S3 on Outposts data transfers between your Outpost and the Region. DataSync allows you to choose what to transfer, when to transfer, and how much bandwidth to use. Backing up your on-premises S3 on Outposts buckets to S3 buckets in the AWS Region allows you to leverage the 99.999999999% (11 9's) of data durability and additional storage tiers (Standard, Infrequent Access, and Glacier) for cost optimization available with the regional S3 service.

Instance replication: You can use CloudEndure to replicate individual instances from on-premises systems to an Outpost, from an Outpost to the Region, from the Region to an Outpost, or from one Outpost to another. The Architecting for DR on AWS Outposts with CloudEndure blog post describes each of these scenarios and how to design a solution with CloudEndure.

Diagram showing disaster recovery (DR) from an Outpost to the Region

Disaster recovery (DR) from an Outpost to the Region

Using AWS Outposts rack as a CloudEndure destination (replication target) requires S3 on Outposts storage.

  • Use EBS snapshots to create point-in-time backups of block storage volumes to Amazon S3 in the Region or S3 on Outposts.

  • Use S3 on Outposts object versioning to maintain multiple versions and history of your objects.

  • Use S3 Replication on Outposts to automatically replicate your object data to another Outpost.

  • For non-data residency use cases, use AWS DataSync to back up objects stored in S3 on Outpost to Amazon S3 in the Region.

  • Use CloudEndure to replicate instances between on-premises systems, logical Outposts, and the Region.