Understanding archive retrieval options - Amazon Simple Storage Service
Restoring large datasetsProvisioned capacity

Understanding archive retrieval options

Amazon S3 has three archival storage classes - S3 Glacier Instant Retrieval, S3 Glacier Flexible Retrieval, and S3 Glacier Deep Archive. While objects stored in the S3 Glacier Instant Retrieval storage class are immediately available using GET, to access data stored in the S3 Glacier Flexible Retrieval or S3 Glacier Deep Archive storage classes you first need to retrieve data using the RestoreObject REST API. Restoring datasets made up of tens of millions of objects or hundreds of terabytes of data could take longer than typical restore times and need special considerations. For more information, see Restoring large datasets.

You can choose from three retrieval access options to restore your archived objects based on your desired retrieval speed – Expedited, Standard and Bulk.

  • Expedited retrieval – Quickly access your data stored in the S3 Glacier Flexible Retrieval storage class or S3 Intelligent-Tiering Archive Access tier. You can use this option for occasional urgent requests for up to hundreds of objects. Objects under 250 megabytes in size are typically made available within 1–5 minutes, and objects 250 megabytes or larger in size are typically retrieved with up to 300 megabytes per second of retrieval throughput. In addition, you have the option to purchase Provisioned Capacity for Expedited retrievals. Provisioned Capacity helps ensure that Expedited retrieval capacity is available when you need it. For more information, see Provisioned capacity.

    Note

    Expedited retrievals are a premium feature and are charged at the Expedited request and retrieval rate. For information about Amazon S3 pricing, see Amazon S3 Pricing.

  • Standard retrieval – Access your data within several hours. Standard is the default option for requests that do not specify the retrieval option. Standard retrievals typically finish within 3–5 hours for the S3 Glacier Flexible Retrieval storage class or S3 Intelligent-Tiering Archive Access tier. Standard retrievals typically finish within 12 hours for the S3 Glacier Deep Archive storage class or S3 Intelligent-Tiering Deep Archive Access tier. Standard retrievals are free for objects stored in the S3 Intelligent-Tiering storage class.

    Note

    • For objects stored in the S3 Glacier Flexible Retrieval storage class or the S3 Intelligent-Tiering Archive Access tier, Standard retrievals initiated by using the S3 Batch Operations restore operation typically start within minutes and finish within 3-5 hours at a throughput of up to 1–2 petabytes per day.

    • For objects in the S3 Glacier Deep Archive storage class or the S3 Intelligent-Tiering Deep Archive Access tier, Standard retrievals initiated by using Batch Operations typically start to complete within 9 hours at a throughput of up to 1–2 petabytes per day.

  • Bulk retrieval – Access your data by using the lowest-cost retrieval option in S3 Glacier storage classes. With Bulk retrievals, you can retrieve large amounts of data inexpensively. For objects stored in the S3 Glacier Flexible Retrieval storage class or the S3 Intelligent-Tiering Archive Access tier, Bulk retrievals typically finish within 5–12 hours. For objects stored in the S3 Glacier Deep Archive storage class or the S3 Intelligent-Tiering Deep Archive Access tier, these retrievals typically finish within 48 hours. Bulk retrievals are free for objects that are stored in the S3 Glacier Flexible Retrieval or S3 Intelligent-Tiering storage classes.

The following table summarizes the archive retrieval options. For information about pricing, see Amazon S3 pricing.

Storage class or tier Expedited Standard (with Batch Operations) Standard (without Batch Operations) Bulk

S3 Glacier Flexible Retrieval or S3 Intelligent-Tiering Archive Access

1–5 minutes

Minutes–5 hours

3–5 hours

5–12 hours

S3 Glacier Deep Archive or S3 Intelligent-Tiering Deep Archive Access

Not available

9-12 hours

Within 12 hours

Within 48 hours

To make an Expedited, Standard, or Bulk retrieval, set the Tier request element in the RestoreObject REST API operation request to the option that you want, or the equivalent in the AWS Command Line Interface (AWS CLI) or AWS SDKs. If you purchased provisioned capacity, all Expedited retrievals are automatically served through your provisioned capacity.

Restoring large datasets

Restoring datasets made up of tens of millions of objects or hundreds of terabytes of data could take longer than typical restore times for any retrieval tier due to retrieval limits.

When you initiate restore requests for objects that are stored in the S3 Glacier Flexible Retrieval, S3 Glacier Deep Archive, or S3 Intelligent-Tiering storage classes, a retrieval-requests quota is applied for your AWS account. S3 Glacier supports restore requests at a rate of 1,000 transactions per second. If this rate is exceeded otherwise valid requests are throttled or rejected and Amazon S3 returns a ThrottlingException error. You can use S3 Batch Operations to retrieve many objects with a single request, which fully utilizes the restore request rate available in your account. For more information, see Performing object operations in bulk with Batch Operations.

After you initiate restore requests, S3 Glacier supports restoring large datasets at a throughput of up to 1–2 petabytes per day per customer account. For any retrieval option, objects larger than 5 terabytes will take longer to be restored with up to 300 megabytes per second of retrieval throughput. For example, a 50-terabyte S3 Glacier Flexible Retrieval object could take up to 48 hours to complete. If you require increased restoration limits, you can contact AWS Support.

Provisioned capacity

Provisioned capacity helps ensure that your retrieval capacity for Expedited retrievals from S3 Glacier Flexible Retrieval is available when you need it. Each unit of capacity provides that at least three Expedited retrievals can be performed every 5 minutes, and it provides up to 300 megabytes per second of retrieval throughput.

Without provisioned capacity, Expedited retrievals might not be accepted during periods of high demand. For predictable, fast access to more of your data, consider using the S3 Glacier Instant Retrieval storage class.

Provisioned capacity units are allocated to an AWS account. Thus, the requester of the Expedited data retrieval should purchase the provisioned capacity unit, not the bucket owner.

You can purchase provisioned capacity by using the Amazon S3 console, the Amazon Glacier console, the Purchase Provisioned Capacity REST API operation, the AWS SDKs, or the AWS CLI. For provisioned capacity pricing information, see Amazon S3 pricing.