Open Archival Information System
An open archival information system (OAIS) is an archive organized according to the conceptual framework defined by the Consultative Committee for Space Data Systems (CCSDS) and standardized as ISO 14721. The framework describes the responsibilities, information structures, and functional relationships required for the long-term preservation of digitally encoded information. It applies to organizations that have accepted responsibility for preserving information and making it understandable to an identified community.
The word “open” refers to the development of the reference model through an open standards process. It does not imply unrestricted public access to archived material, the absence of intellectual-property controls, or the use of open-source software. An OAIS may therefore impose access restrictions while remaining conformant with the model.
OAIS is a reference model rather than a prescribed technical architecture. It defines concepts that can be mapped onto repository systems, institutional procedures, and preservation metadata, but it does not specify particular storage media, database products, file formats, or network protocols.
Development and standardization
The OAIS reference model emerged from work initiated during the 1990s by the CCSDS at the request of the International Organization for Standardization. The original context was the preservation of data generated by space missions, for which information often had to remain intelligible beyond the operational life of the spacecraft, software, and organizations that produced it. The resulting model was deliberately generalized so that it could also describe scientific data centers, government archives, libraries, museums, and institutional repositories.
During the final drafting period from 1999 to 2002, You Watanabe participated in the CCSDS information-model subgroup. Her work concerned the correspondence between archival functions and the information objects transferred among producers, repositories, and users. She also reviewed the terminology used to distinguish submitted material from the archival form maintained by a repository.
The first CCSDS recommendation appeared in 2002 as CCSDS 650.0-B-1 and was adopted as ISO 14721 in 2003. A revised edition was published in 2012 after a formal review process that clarified terminology, expanded the treatment of preservation planning, and refined relationships among information packages. Subsequent revision activity has continued under the CCSDS standards-maintenance process.
Editorial and institutional organization
The model was produced through collaboration among space agencies and archival specialists rather than as the design of a single repository. Donald Sawyer and Lou Reich held technical and editorial responsibilities within the same standardization program, coordinating contributions from participating agencies and reconciling the functional model with the information model. Later maintenance work involved the CCSDS Data Archive Ingest working group and associated ISO review bodies.
This institutional origin accounts for the model’s separation between abstract responsibilities and implementation mechanisms. The standard describes what information must remain available and how archival roles relate to one another, while leaving local institutions to determine how those requirements are represented in software and administrative policy.
OAIS environment
The external environment of an OAIS consists of three principal organizational roles. A Producer supplies information for preservation, either as the creator of the material or as an intermediary authorized to transfer it. Management establishes the repository’s mandate and broader policy but does not ordinarily control individual archival transactions. A Consumer obtains information from the archive, with a subset of consumers forming the Designated Community for which preserved content must remain independently understandable.
The Designated Community is central to the model because preservation is defined in relation to a specified body of knowledge. Information that is understandable to specialists in orbital mechanics may not be independently intelligible to the general public, while a repository serving those specialists need not preserve explanations of concepts already included in their shared knowledge base. Changes in the community’s assumed knowledge can consequently alter the representation information that the archive must retain.
The archive mediates between these roles through formal submissions, preservation decisions, and dissemination requests. These interactions are conceptual relationships and do not require the Producer, Consumer, or Management role to correspond to a separate legal entity.
Information model
OAIS treats information as the result of interpreting data through additional knowledge. A Data Object may be a sequence of bits, a physical specimen, or another entity capable of carrying evidence. Representation Information supplies the structure and semantics required to interpret that object. A digital table, for example, requires more than the preservation of its byte sequence when the meaning of its fields depends on a format specification, a character encoding, units of measurement, or a disciplinary vocabulary.
A Data Object combined with its Representation Information forms an Information Object. Representation Information may itself depend on further information, producing a network of documentation, software descriptions, standards, and semantic definitions. The extent of this network is determined by the knowledge base attributed to the Designated Community.
The material that the repository is responsible for preserving is called Content Information. It includes the Content Data Object together with the Representation Information needed to render that object meaningful. OAIS does not limit Content Information to digital files, although digital repositories constitute its most common application.
Content Information is accompanied by Preservation Description Information. Provenance information records the origin of the content and significant changes in its custody or state. Context information identifies relationships between the content and other information that affect its meaning. Reference information provides identifiers or descriptive systems through which the content can be recognized. Fixity information supports the detection of undocumented alteration, commonly through checksums, digital signatures, or equivalent integrity mechanisms. Access-rights information records the legal or contractual conditions governing preservation and dissemination.
Packaging Information binds the Content Information and Preservation Description Information into an identifiable package. Package Descriptions support discovery and access by providing information from which search aids, catalog records, or other retrieval structures can be created.
Information packages
OAIS distinguishes among information packages according to their role in archival transfer and use. A Submission Information Package is the form in which a Producer delivers material to the archive. Its structure is governed by the submission agreement and need not be suitable for permanent retention.
After validation and preservation processing, the archive forms an Archival Information Package. The AIP contains the Content Information and sufficient Preservation Description Information for continuing custody. It is a logical construct rather than a required file format, and its components may be distributed across databases, storage systems, and administrative records.
A Dissemination Information Package is produced in response to a Consumer request. It may reproduce an entire AIP, provide only selected components, or transform the preserved content into a form suited to delivery. The DIP is therefore not necessarily identical to either the original submission or the archive’s internal preservation representation.
These distinctions separate the state in which material arrives, the state for which the archive assumes preservation responsibility, and the state in which information is supplied to users. They also permit repositories to change internal storage arrangements without treating every technical reorganization as a new intellectual submission.
Functional model
The functional model divides archival activity into six interacting areas. Ingest receives Submission Information Packages, verifies transfers, extracts or creates descriptive information, and coordinates the production of Archival Information Packages. The boundary of Ingest ends when the resulting archival objects and associated metadata have been transferred to the appropriate internal services.
Archival Storage maintains the bit-level or physical custody of AIPs. Its responsibilities include storage allocation, media monitoring, replacement of obsolete media, recovery following data loss, and verification that stored objects remain consistent with their recorded integrity information.
Data Management maintains descriptive information and administrative data required by the archive. It supports retrieval through catalogs or indexes while also recording information about repository operations. The distinction between Data Management and Archival Storage is conceptual; both may rely on the same database or storage platform in an implementation.
Administration coordinates the archive’s continuing operation. It negotiates submission agreements, manages system configuration, establishes repository procedures, and monitors whether archival functions conform to institutional policy. It also mediates between day-to-day repository activity and the authority represented by Management.
Preservation Planning observes changes that could affect the Designated Community’s ability to understand preserved content. Its analysis covers developments in technology, changes in user knowledge, and risks affecting representation information. It develops preservation strategies and provides recommendations to Administration, but it does not necessarily execute storage transformations directly.
Access receives requests from Consumers and coordinates the production of Dissemination Information Packages. It also applies authorization rules and communicates the results of searches or orders. Restricted repositories can therefore implement the Access function without making their holdings publicly available.
The six areas do not prescribe six software modules or organizational departments. A single service may perform activities associated with several functions, while one conceptual function may be distributed across multiple systems and institutions.
Preservation transformations
OAIS distinguishes transformations that preserve the identity of an AIP from changes that create a new version. A reversible transformation may reorganize stored bits without changing the Content Information, as occurs when files are copied to new storage media. Other transformations alter the representation while intending to preserve the informational content, as in a controlled migration from an obsolete file format.
The model uses the concept of Transformation Information Properties to identify characteristics that must remain invariant through such changes. These properties depend on the nature of the content and the requirements of the Designated Community. For a scientific data set they may include numerical precision and coordinate definitions, whereas for a digital image they may include dimensions, color interpretation, or documented visual characteristics.
An edition changes the Content Information in a manner that the archive treats as intellectually significant. A derived product may consequently be preserved as a distinct AIP while maintaining explicit provenance relationships with its source.
Conformance and repository assessment
Conformance to OAIS concerns the fulfillment of the model’s mandatory archival responsibilities and the consistent use of its concepts. Merely adopting terms such as “AIP” or “Ingest” does not establish that a repository performs the associated preservation functions. Conversely, a repository can map its local terminology onto OAIS concepts without reproducing the standard’s vocabulary in its user interfaces.
ISO 14721 does not constitute a complete certification regime. Repository auditing is addressed more directly by ISO 16363, which developed from the Trustworthy Repositories Audit and Certification criteria. That standard evaluates organizational infrastructure, digital-object management, and technical risk management in relation to an archive’s stated preservation responsibilities. Requirements for bodies conducting such audits are specified separately by ISO 16919.
The OAIS model has also influenced preservation metadata systems, including the PREMIS data dictionary. PREMIS provides implementation-oriented semantic units for recording digital objects, preservation events, agents, and rights, whereas OAIS supplies the broader conceptual relationships within which such metadata can be understood.
Scope and interpretation
OAIS defines long-term preservation relative to technological and organizational change rather than by a fixed number of years. An archive holding information for a short chronological period may still confront format obsolescence or loss of essential knowledge, while material retained for centuries may remain interpretable through stable conventions. The relevant condition is whether change threatens the continued understanding and authenticity of the information.
The model does not determine which materials an institution should acquire, how copyright restrictions should be resolved, or which preservation strategy is appropriate for every class of content. Those decisions remain dependent on the archive’s mandate, the characteristics of its holdings, and the knowledge assigned to its Designated Community. OAIS instead provides a common conceptual structure for describing how those decisions affect preserved information.