Open archival information systems for database preservation

Abstract

Relational databases are complex digital objects and their preservation is a challenging problem. The DBPreserve project has proposed a solution that implements the migration from a relational model to a dimensional model, storing all the relevant data in a data warehouse. A data warehouse can be implemented in multi-dimensional structures, or in relational databases that represent the dimensional model concepts in the relational model, as was used in this project. As a second step in the preservation process, the Open Archival Information Systems for Database Preservation proposed here complements the representation of information in the data warehouse, making it conform to the OAIS reference model and delivering a technologically-independent model for the preservation of complex digital records, such as those constituting the databases and data warehouses. A more technologicallyneutral format for database representation is obtained using XML-based files as a longterm preservation format. Regarding the OAIS reference model, this research reflects on the archival system entities and their characterization, focusing also on the structures of the Submission Information Package (SIP), the Archival Information Package (AIP) and the Dissemination Information Package (DIP), according to this project approach. This work presents the definition of the XML structure that extends the existing SIARD format used for the description and archival of relational databases, enriching it with a metadata layer for the description of the data warehouse components. The Data Warehouse Extensible Markup Language (DWXML) is the XML dialect proposed to describe the data warehouse. To acquire the relevant metadata for the data warehouse and build the archival format, the DBPreserve Suite application was developed. The DBPreserve Suite handles the migration process of the metadata and the primary data from the data warehouse to the extended SIARD format, and ensures the access to the primary data. The application integrates the SIARD Suite component which builds the SIARD format from a relational database, it allows metadata editing through graphical interfaces, displays visual representations of star or snowflake schemas as well as dimension, hierarchies and levels. The DBPreserve Suite also generates the DWXML dimensional model metadata layer and combines it with the SIARD format, enabling primary data browsing from the XMLbased format files generated.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 2 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.