A digital archive remains trustworthy and usable only when its stewards manage more than storage. They need a defined scope, protected copies, documented changes, bit-level checks, and a plan for keeping files readable. Start with the records’ value, users, rights and likely threats; then choose controls that fit the collection and its resources. A checksum can reveal that file contents differ from a trusted baseline, but it cannot explain why, prove who created the file, or show whether people can still open and interpret it.
What does archive integrity mean?
Archive integrity is the continued reliability of records as a collection and as individual files. It involves knowing what was received, keeping the files complete, recording custody and preservation actions, controlling who can change or view material, and making the records usable for their intended audience. These aims overlap, but no single check proves them all.
Fixity is a bit-level check: a checksum calculated from a file is compared with a trusted reference value. A match indicates that the compared contents are the same; a mismatch shows that they differ. The checksum does not locate the changed bits or tell you whether the change was corruption, an authorized edit, a transfer error or tampering. It also cannot establish a file’s creator or truth. The Digital Preservation Coalition’s fixity and checksum guidance describes checks at receipt, in storage and when files are delivered. For the broader distinction between integrity and authorship, see content authenticity and verification tools.
| Control | What it helps answer | What it does not establish on its own |
|---|---|---|
| Checksum comparison | Did these file contents change from the trusted reference? | Who changed them, why, or whether the reference is trustworthy |
| Inventory and metadata | What files belong in the collection, and what are they? | That every file is readable or unchanged |
| Provenance and event records | Where did material come from, and what actions were recorded? | That an incomplete or unprotected log is a complete account |
| Access and format tests | Can intended users find, open and interpret the material? | That its bits have remained unchanged |
How to preserve digital artifacts: a practical checklist
- Define the collection and its purpose. Make a manageable inventory: what material you hold, where it came from, why it must be retained, who is responsible, who may use it, and what systems or formats it depends on. Prioritize records by value, sensitivity, retention needs and the consequences of loss. The Digital Preservation Coalition’s getting-started guidance recommends beginning with a rapid assessment of the organization, its goals and the materials it needs to preserve.
- Set policy and responsibilities. Write down which material is in scope, how it is acquired and described, who can access or change it, how preservation events are recorded, how recovery decisions are made, and who approves migrations or deletion. Include rights and restrictions that apply to the collection. Assign an owner for each task and a way to review the policy when the collection, risks, technology or resources change.
- Establish trusted records at ingest. Check that expected files arrived and that a supplied manifest matches the deposit. Where the supplier provides a checksum from a trusted channel, compare it at receipt. If there is no independent reference, calculate a baseline during a controlled ingest and record the algorithm, value, file identifier, date and action. A checksum stored beside a file is useful only if the reference is protected from unauthorized replacement too.
- Make recovery possible. Keep more than one usable copy when the collection’s value and risks warrant it, and consider whether copies share the same location, administrator, credentials or failure mode. There is no copy count that fits every archive: choose redundancy based on threats such as accidental deletion, equipment failure, site-level disasters or malicious changes, plus the recovery time and cost the organization can support. Separate backup and preservation needs, protect copies with appropriate access controls, and test that a recovery actually works.
- Check transfers and stored files. Compare checksums after a copy or move, and compare stored files with their trusted references as part of a documented monitoring process. Investigate mismatches rather than assuming malice or corruption. Review event logs and other copies; if a known-good copy is available, restore from it and verify the restored file. Record the finding, decision and repair. Both guides make check frequency context-dependent. The DPC also cites use and processing resources; the University of Melbourne guide also names collection size, costs and environmental impact. Choose an interval for your own system instead of applying a universal schedule.
How should a preservation policy guide daily work?
A policy is useful when it turns broad goals into repeatable decisions. State what the archive is meant to preserve and for whom, who is accountable, what access is permitted, how original material is protected, what documentation must accompany it, and what to do when a check fails. Make the procedure clear enough that another authorized staff member can understand the collection and continue the work.
Include the records needed to interpret later actions: file and format identifiers, source and custody information, checksums and their algorithms, access restrictions, and preservation events such as a transfer, fixity check, restoration or format conversion. The Library of Congress describes preservation metadata as including information for reference, context, fixity and provenance; its PREMIS resources provide a preservation-metadata standard. Use a standard or repository feature that fits your workflow rather than collecting fields nobody can maintain.
Review the policy when a material change could affect the archive—for example, a storage provider or format is no longer supported, a new risk emerges, the audience changes, or a recovery test fails. The National Digital Stewardship Alliance’s Levels of Digital Preservation can help teams assess and improve practices. Version 2.1, released in March 2026, adds environmental-sustainability guidance; the NDSA says the highest level is not necessarily the goal for every organization or every area.
How should an archive choose and protect storage copies?
Choose storage by the material, access requirements and threats—not by a slogan about a specific device. A copy that is easy to reach may help routine work; a separate copy may help recovery from a local failure. Think through who controls each copy, how it is protected from accidental or unauthorized changes, how quickly it can be restored, and whether staff can verify it. A second copy under the same permissions and in the same failure domain may not address the risk you are trying to reduce.
Backups support recovery, but a backup alone does not prove that files are complete, authentic or still readable. Pair copying with an inventory, fixity records, access controls and a tested restoration process. When a checksum mismatch occurs, do not overwrite the file immediately: first preserve the evidence and compare available copies and event records so the recovery decision is documented.
Algorithm choice also depends on the threat. The DPC says MD5 can be sufficient for detecting accidental loss or damage, while a stronger algorithm such as SHA-256 is appropriate when resistance to deliberate alteration matters. Do not treat MD5 as a universal security choice, and do not assume a stronger digest proves authenticity if an attacker can replace both the file and its reference. For a plain-language explanation of the process, see how checksum algorithms protect digital content. A digital fingerprint for content verification is one part of a documented verification process, not a substitute for secure custody and provenance.
When should files be migrated, and how do you keep them accessible?
Storage media and file formats create different risks. A storage transfer moves bits to another medium; a format migration changes how information is encoded so it can continue to be rendered or used. Neither should be triggered by an arbitrary calendar alone. Plan a migration when evidence—such as a failing dependency, an unsupported format, a change in user needs or an inability to open important files—shows that access is at risk.
Before choosing a target format, consider the content’s important features, documentation, adoption, transparency, dependencies and the tools available to validate it. The Library of Congress’s format sustainability factors explain why “open” by itself is not a complete test: the right choice also depends on the kind of material and what future users need to see or do with it. Retain the original when policy and rights allow, test representative output, document the source and result, note the tool and decisions used, and verify the migrated files. Because a migration intentionally changes file contents, calculate and record a new fixity baseline for the output.
Access also depends on context. Preserve enough descriptive and technical metadata for people to find and interpret the material, document relevant software or dependencies, and test that files open for the intended users under the access rules. Persistent identifiers can help people refer to an item as its location changes, but an identifier does not preserve the file or guarantee that it remains available; see this related guide to persistent identifiers and copyright protection.
What can a fixity check prove in practice?
Suppose an archive receives a folder of meeting records and moves it into preservation storage. Staff check the file list, compare any trusted checksums supplied with the deposit, and record a baseline for files without one. After transfer, they compare the stored files with those references. If one value differs, they check the transfer and preservation logs and compare other controlled copies before deciding whether to restore the file. They record what they found and verify the replacement.
If a later format migration is needed so staff can open a record, the archive documents why the conversion was made, tests the output, and creates a new checksum for that version. The original and converted file are not bit-for-bit identical, so the old checksum is not a valid expected value for the new file. The comparison supports a claim about bit-level continuity; provenance, reliable records of actions and access testing are still needed to explain the record and its usability.
When should an institution consider a shared preservation service?
Some organizations may need a community service or preservation network rather than building every capability alone. APTrust’s current FAQ describes a community-governed repository for eligible institutions and explains its membership and preservation services. The LOCKSS Program describes a different model: distributed preservation networks, including a Global LOCKSS Network focused on scholarly journals and books and local access. These services have different eligibility, content scope, operating models and costs; compare their current terms, access, export and recovery arrangements with your collection’s needs before relying on one.
For a task-by-task comparison of archive fixity software and workflows, see digital archive fixity tools.

