Don't Panic — Data Archiving

Data archiving is the disciplined act of moving selected information out of routine use without turning it into a box of mysterious old files. The box is cheap. Knowing why an object is there, who may retrieve it, and what it means years later is the expensive bit. Storage is delighted to keep bits. It has no strong opinions about whether anyone can use them.

An archive starts with a purpose. Perhaps a record must support an audit, a legal discovery request, research, or historical preservation. That purpose creates a retention rule: keep this class of information for this period, then destroy it, transfer it, or continue preservation. Before an archive, old material tends to remain in active systems or get mistaken for a backup. Both arrangements eventually acquire a personality, and it is rarely a helpful one.

The key object is an archive package: content plus the context that makes it usable. A database export without its schema, data dictionary, encoding, and relationship rules is a confident-looking pile of tables. Metadata supplies identity, owner, retention rule, format, rights, checksum, and the history of ingest or migration. The archive is keeping an explanation alongside the bytes because future readers are not telepathic, however unfair that seems to the bytes.

Fixity is the archive's answer to a narrow but vital question: did these checked bits change unexpectedly? Compute a checksum at ingest, keep it, and compare it during later checks. A match supports bit-level integrity; it does not prove the original was correct, locate damage, repair it, or make an obsolete format readable. That is why independent copies, format information, and documented recovery procedures are not decorative extras.

The surprise is that a successful copy is not proof of a successful archive. Retrieval is the test. Can an authorized requester find the correct object, obtain every component, verify it, decrypt it when necessary, and interpret it with retained documentation? If any answer is no, the archive has preserved a future problem with admirable reliability.

Read the Intro for the full lifecycle and its decision points. Use the Slides for the map from selection through disposition. Keep the Cheatsheet nearby when designing metadata, fixity checks, copies, migrations, and retrieval tests. Then use the Reference tab when a policy or implementation choice needs authoritative detail. The task is not to keep data forever. It is to keep the right information usable for the right reason, and to know what happens next.

Don't Panic — Data Archiving

Relevant careers