3 Digital preservation storage has a time horizon that few other IT systems share. National libraries, national and regional archives, research institutions and cultural heritage organizations are expected to keep digital collections usable not for years but for generations. Web archives, digitized books and newspapers, born-digital government records, audiovisual collections and research data all need to survive hardware failures, technology changes, budget cycles and the slow decay of storage media, while remaining authentic and accessible. This article explains what digital preservation asks of storage, the architectures institutions use and how to plan a preservation storage platform. It is written for digital preservation managers and IT architects in libraries, archives and similar organizations. What makes preservation different Most IT storage is designed around applications and business needs that change every few years. Preservation storage is designed around content that must outlive many generations of hardware and software. That leads to a distinct set of priorities: Integrity over time: content must be bit-for-bit identical decades later, with evidence that it has not changed. Multiple independent copies: so no single failure, disaster or mistake can destroy the collection. Technology independence: content must be movable to new storage without loss, again and again. Authenticity: provenance and changes must be documented. Accessibility: content must be findable and usable by future researchers and the public. Sustainability: costs must be affordable year after year, since preservation never ends. Frameworks that shape preservation storage OAIS The Open Archival Information System reference model, published as an international standard (ISO 14721), describes the functions and responsibilities of an archive that preserves information for a designated community. It defines how content is ingested, stored, managed and made available, and it is widely used to design and assess preservation systems. Storage falls mainly under its archival storage function. See what the OAIS model asks of storage. NDSA Levels of Digital Preservation The National Digital Stewardship Alliance in the United States publishes the Levels of Digital Preservation, a practical tiered guide covering storage, integrity, control, metadata and content. Its storage guidance progresses from keeping two complete copies in separate locations toward keeping more copies in geographically separate locations with different disaster threats. Many institutions outside the US also use it as a self-assessment tool. Trustworthy repository certification Certification schemes such as CoreTrustSeal and the ISO 16363 audit standard assess whether repositories are trustworthy, including how they manage storage, integrity and risk. What preservation storage has to deliver Multiple copies in separate places Preservation practice calls for several copies of every item, stored in different locations with different risks, often on different technologies. See how many copies digital preservation needs, and where. Fixity checking Fixity means verifying that content has not changed, usually by computing checksums and comparing them over time. At petabyte scale, regular fixity checking is a significant workload. See how archives run fixity checks at petabyte scale. Media and technology refresh Drives fail, tape formats age and storage systems reach end of support. Preservation storage must allow content to move to new hardware regularly without loss, ideally without disruptive migration projects. Long-term cost sustainability Preservation budgets must cover storage for the indefinite future. Cost per terabyte per year, including power, space, staffing and refresh, matters more than purchase price. Access Some collections are accessed heavily, such as popular digitized collections or web archives; others are rarely touched. Storage should support access needs without forcing everything onto expensive fast storage. Security and control Collections may include sensitive government records or personal data. Access must be controlled and audited. Content must also be protected from ransomware and malicious deletion. Storage architectures for preservation Tape Tape libraries have long been central to preservation for their low cost per terabyte, low power use and offline protection. Tape generations change regularly, so content must be migrated to new formats, and access is slower than disk. Disk-based object storage Object storage on standard servers offers online access, scale-out growth, erasure coding, integrity checking and the ability to replace hardware node by node. Many preservation systems and repository platforms can use S3-compatible storage. See object storage vs traditional storage. Cloud storage Some institutions use cloud storage for one of their copies. Considerations include long-term cost, retrieval charges, data sovereignty and the institution’s ability to move content out later. Combined approaches Most national institutions combine technologies: an online object storage copy for access and fixity checking, a second copy at another site and often a tape or offline copy for additional protection against correlated failures and cyberattacks. Sovereignty and legal deposit National libraries and archives typically operate under legal deposit and public records laws that make them responsible for national collections. Many prefer, or are required, to keep preservation copies within national borders and under direct control. In Europe, national institutions often keep all copies in-country. In Japan, the UK, France and Germany, similar expectations apply to national collections. Storage that can be deployed across multiple domestic sites, without dependence on foreign cloud regions, suits these requirements. See data sovereignty requirements. Integrity and authenticity Preservation storage must support integrity at every stage: Checksums computed at ingest and stored as preservation metadata. Regular fixity verification across all copies. Repair from good copies when corruption is detected. Audit logs of storage operations. Immutability for content that must never change, protecting against accidental or malicious deletion. See S3 object lock: immutability and WORM. Object storage platforms that verify checksums in the background add a layer of protection beneath the preservation system’s own fixity checks. See data durability in high-density storage systems. Collections with different needs Not every collection behaves the same way, and preservation storage should reflect that: Web archives grow quickly through regular national domain crawls, produce very large numbers of files packaged in container formats and are heavily used by researchers once access is opened. Digitized print collections such as books, newspapers and maps generate high-resolution master images plus derivatives for online access, with strong public demand for popular titles. Audiovisual collections are among the largest by volume, with uncompressed or lightly compressed masters from digitized tape and film that can run to gigabytes per hour of content. Born-digital records transferred from government departments arrive in many formats, often with sensitivity restrictions that limit access for years. Research data may need preservation for reproducibility, with variable formats and sizes. Grouping collections by size, growth, access and sensitivity helps decide which ones need online access copies, which can rely mainly on deep storage and how fixity schedules should be set. Staffing and skills Preservation storage succeeds or fails on the people who run it over decades. Institutions need staff who understand both preservation principles and storage operations, documented procedures that survive staff turnover and relationships with suppliers that can support long-lived platforms. Choosing storage that is simple to operate and expand, using standard hardware and open interfaces such as S3, reduces dependence on scarce specialist skills and makes it easier to change suppliers if needed. Planning a preservation storage platform Define collections, volumes, growth and access patterns. Map requirements from OAIS, the NDSA levels and certification goals. Decide on the number and location of copies and the technologies for each. Plan fixity checking frequency and capacity. Choose storage that allows hardware refresh without content migration projects. Model cost over at least ten years. Integrate storage with repository and preservation systems. Document everything for audit and succession. Checklist: digital preservation storage Keep multiple copies in separate locations with different risks. Use at least two storage technologies where possible. Compute and store checksums at ingest. Run regular fixity checks and repair from good copies. Protect content with immutability and access control. Plan for media and hardware refresh without loss. Keep copies within required national boundaries. Model long-term cost sustainability. Align with OAIS and certification requirements. Putting it together Digital preservation storage is built for permanence: multiple independent copies, constant integrity checking, technology refresh without loss and costs that remain sustainable indefinitely. National libraries and archives typically combine online object storage, a second site and an additional copy on a different technology, all kept under national control. Choose platforms that make refresh routine rather than disruptive, and design every layer so it can be audited and handed on to the next generation of stewards. Frequently asked questions What storage do national libraries use? Typically a combination: online disk or object storage for access and fixity checking, additional copies at separate sites and often tape or offline copies. How many copies does digital preservation need? Preservation guidance generally calls for at least two, and often three or more, copies in separate locations with different risks. What is fixity in digital preservation? Verifying that content has not changed, usually by comparing checksums computed at ingest with checksums computed later. Is cloud storage acceptable for digital preservation? It can be used for a copy, but long-term cost, retrieval charges, sovereignty and exit options need careful assessment. What is OAIS? The Open Archival Information System reference model, an international standard describing the functions and responsibilities of a long-term digital archive. Further reading See the OAIS model and storage, digital preservation copies, fixity checks at scale, research data management storage and data archiving best practices.
