The role of DPS in the NLN

The Digital Preservation Services (DPS) is the National Library’s system for long-term digital preservation. The DPS does not operate in isolation: it is one of three system domains that share responsibility for digital objects at the National Library. These domains operate independently, but each domain’s responsibilities influence the others. This page describes the context the DPS was built in and how that context shapes the system.

The three system domains described here reflect the National Library’s internal operational landscape.

External clients interact with the DPS directly through the API, bringing their own identifiers, metadata, and scope definitions.

The three system domains

Within the National Library’s own collections, responsibility for any digital object¹ is spread between three system domains, each managing different aspects of the object:

Metadata management systems

Catalogs, registries, and collection management systems (such as Alma, Mavis, and Hanske). They are the authoritative source for descriptive metadata and issue unique identifiers (UIDs, typically URNs) for digital objects. These systems manage a wide range of entities; only a subset describe digital objects relevant to digital preservation.

This is where intellectual entities are defined: the choice of what constitutes a preservable information package originates here, not in DPS. A single URN typically maps to a single package. For details on scoping, see Package, representation and data scope.

Digital Preservation Services (DPS)

The preservation system. Manages preservation data using a PREMIS-based data model (Intellectual Entities, Representations, Files, Events, Agents). The DPS provides “cold” long-term storage with asynchronous access, designed for preservation rather than immediate retrieval. The DPS is the authoritative source for technical file metadata in the National Library. It is not the authoritative source for descriptive metadata: it keeps the metadata it needs to operate independently, without replicating the full cataloguing structures of the metadata management systems.

For external producers and consumers, the DPS operates as a self-contained service: you provide your own identifiers and metadata through the API, and the DPS manages preservation independently. For details on how data and metadata are handled within DPS, see Data management and Metadata management.

Public access services

Nettbiblioteket and Oria. Manage a harvested subset of descriptive metadata and provide access copies derived from preservation files. These services offer “hot” storage with immediate access, unlike the preservation storage of the DPS.

The public access services transform the harvested metadata into a flattened structure, operating with a single representation per intellectual entity. As a result, the intellectual entities found online do not necessarily mirror the entities in the metadata management systems.

A UID, managed by the metadata management systems, ties the three domains together. It must be unique across systems and should not be reused.

Data and metadata flow

The diagram below is an idealized and simplified view of how data and metadata move between the domains:

The SIP carries preservation data and a copy of descriptive metadata from the metadata management systems. The DPS does not preserve access copies that can be automatically derived from preservation files; such copies are managed by the public access services.

PREMIS entities and overlapping responsibilities

The DPS uses PREMIS as its preservation metadata model. Neither the metadata management systems nor the public access services use PREMIS natively, but PREMIS concepts help explain where responsibilities begin and end across the three domains within the NLN. This mapping explains why the DPS was designed the way it is, not how external users should model their own data².

Within the NLN, PREMIS entities are distributed across the three system domains. For external users, the DPS manages its own data model (Intellectual Entities, Representations, and Files) and scope is defined by the client. The roles of the other two domains are filled by your own systems.

• Metadata management systems: within the NLN, this is where intellectual entities are defined and identified. The representation level is typically not described in these systems. They define representations implicitly through the representation’s 1:1 relationship with the IE.
• Digital Preservation Services (DPS): manages the full PREMIS data model: Intellectual Entities, Representations, and Files. The DPS does not replicate the full descriptive metadata structures of the metadata management systems. It offers searches based on the metadata provided to and generated by the DPS. For details, see Metadata management.
• Public access services: manage and disseminate a harvested subset of intellectual entities, access representations, and files.