Glossary
Storage lifecycle management (SLM)
Storage lifecycle management (SLM) is the policy-driven handling of data and the storage beneath it from creation to deletion. Rules based on age, access, value and retention obligations decide where data sits, how it is protected, when it moves and when it is removed, so the decisions are made once rather than file by file.
The stages data passes through
- Creation and ingest. Data arrives and is classified, often by tags or metadata, so later rules know what it is.
- Active use. Data is read and changed often and sits on fast media such as hot storage.
- Infrequent access. Reads drop off and the data moves to cheaper capacity that stays online.
- Retention and archive. Data is kept for legal, regulatory or reference reasons, often in cold storage, and may be locked against change.
- Expiry and deletion. Once retention ends and no hold applies, the data is deleted and its space returns to the pool.
Not every dataset passes through every stage. Training data for an AI model can return from archive to active use when a model is retrained, and model checkpoints can go from active to deleted in weeks.
What can a lifecycle policy do?
A policy matches data by attributes such as age, last access, location or tags, and then acts on it. Typical actions are moving it to another tier, changing its protection level, replicating it to another site, locking it for a retention period, and deleting it when that period ends. One rule written for a class of data applies to billions of objects without anyone touching them individually.
SLM, tiering and data lifecycle management
| Term | Scope | Typical owner |
|---|---|---|
| Storage tiering | Placing data on media by how often it is used | Storage team |
| Storage lifecycle management | Placement plus protection, replication, retention, immutability and deletion | Storage and infrastructure teams |
| Data lifecycle management | The full useful life of data, including governance and business value | Data owners, compliance and IT together |
Tiering is one action inside SLM, and SLM is the infrastructure side of data lifecycle management.
Protection has to follow the stage
Protection is not constant across a dataset's life. Active data may be replicated across sites for availability, while archived data is kept on efficient erasure-coded capacity and locked against change. Data under a retention obligation must not be removable by an age rule alone, so retention locks and legal holds sit above deletion rules. Rules that delete are where errors cost most, which is why storage automation practices such as staged rollout apply here directly.
Why object storage suits it
Each object carries its own metadata, and policies can act on it without a separate catalogue. That lets object storage apply lifecycle rules across very large, unstructured estates, where reviewing data by hand is impossible.
Frequently asked questions
Who sets storage lifecycle policies?
Retention periods come from data owners, legal and compliance. The storage or infrastructure team turns them into rules and decides placement and protection for each stage.
How is storage lifecycle management different from storage tiering?
Tiering only decides which media data sits on. Lifecycle management also covers protection, replication, retention, immutability and deletion over the data's whole life.
How often are lifecycle policies reviewed?
Usually whenever retention requirements, access patterns or storage costs change, and at least once a year, because a rule written for last year's data can move or delete the wrong data today.