Amazon Glacier is a service optimized to store data with low cost and for long-term retention, it provides the same durability as S3, that is, 119s and security at rest with AES256 for all objects.

Glacier is a cold storage solution meaning that access is not on real time. If data availability is required, consider the use of the Simple Storage Service. Amazon Glacier is a great option to achieve the lowest cost for storing cold data.

Data in Glacier is aggregated via archives. Archives can be anything like a document, video, a ZIP file, and so on. This archives can hold up to a maximum of 40 TB. Archives are stored in vaults and every account has a limit of 1000 vaults.

Information can be retrieved through retrieval jobs. Data archiving can be reinforced with Vault Lock Policies that can be used to meet storage retention and compliance, and to achieve write once read many (WORM) storage capabilities.

When data is restored, Glacier will copy the object to S3 so you can have on-demand access to the data. If you configured the notification topic you will receive an SNS notification when the data is retrieved. If data stored on Glacier was managed by S3 this data needs to be restored using the S3 APIs, on either case you’ll have to retrieve it using the Glacier APIs.