Data archiving preserves information that is no longer needed for routine operations but still has business value or must be retained. Examples include completed project files, historical application logs, financial statements, research datasets, and original media assets. An archive needs more than stored bytes: engineers must also preserve enough metadata, ownership information, and format documentation to locate and interpret the data later.
Object storage is a common cloud foundation for archiving. It stores data as objects in buckets or containers, with identifiers and metadata, rather than exposing disk blocks to a server. Its scalability and support for large quantities of unstructured data make it well suited to documents, images, exports, and log files. Unlike block storage used by an active database, an archive usually does not need frequent, low-latency updates. Object storage is not automatically archival, however; the selected storage tier and operating policies determine its suitability.
A practical implementation classifies data by retention requirements, access frequency, and acceptable retrieval time. Lifecycle policies can transition eligible objects from frequently accessed storage into lower-cost tiers and eventually expire data when permitted. For example, an organization might keep application logs readily accessible for troubleshooting, move older logs into an archival tier, and delete them after its approved retention period. Legal holds or mandatory retention controls must prevent deletion when an investigation or regulation requires continued preservation.
The main economic trade-off is storage cost versus access cost and speed. Some archival tiers provide immediate access, while others require a restore operation and a wait that can range from minutes to hours. Retrieval charges, request charges, minimum storage durations, and charges for network transfer can change the overall cost. Thousands of tiny objects or frequent retrievals can make the cheapest advertised storage rate a poor fit. Test representative retrievals and calculate total costs before selecting a tier.
On AWS, Amazon S3 is the core object storage service relevant to this use case. Its archival options include S3 Glacier Instant Retrieval, S3 Glacier Flexible Retrieval, and S3 Glacier Deep Archive. These are storage classes with different access characteristics, not interchangeable choices. Instant Retrieval supports millisecond access, whereas Flexible Retrieval and Deep Archive require restoration before object data can be read. Choose according to the required retrieval deadline rather than assuming every Glacier class behaves identically.
Archiving is also distinct from backup. Backups primarily support recovery after corruption, deletion, or system failure; archives primarily support long-term retention and later reference. Either can use object storage, and archival tiers can hold eligible backup data. Neither high durability nor low storage cost automatically provides immutability, regulatory compliance, or protection against authorized deletion. Apply appropriate access controls, encryption, retention locks, integrity checks, and retrieval tests, and ensure encryption keys remain available for the archive's lifetime.
Amazon S3