Glossary Term

Data Tiering

What is data tiering?

Classifying and moving data across storage tiers based on how often it's accessed, how old it is, and what policies apply, so you spend less without giving up performance.

Data tiering splits files across storage levels, usually called hot, warm, and cold. Data that people touch regularly (hot) lives on fast storage. Files nobody has opened in months (cold) get moved to cheaper options like cloud archives, object stores, or tape. The idea is simple: match what you’re paying to store something with how much that something actually matters to your work.

When you’re managing billions of files, nobody is sorting this by hand. A metadata-driven approach handles it by looking at file attributes like last access and modified time, file type, and ownership. Policies built on that metadata promote or demote data between tiers automatically, so IT teams aren’t stuck fielding tickets about what goes where.

Tiering also helps deal with data ROT (Redundant, Obsolete, and Trivial files), which tends to pile up and eat expensive capacity. Moving that inactive data to cheaper tiers frees up fast storage for the work that actually needs it, whether that’s AI/ML data preparation or real-time research computing.