Last updated:

July 4, 2019

What is the SIARD research project?

In 2013-14, Public Record Office Victoria (PROV) evaluated the Software Independent Archiving of Relational Databases (SIARD) tools and format. The SIARD research project is part of PROV’s ongoing work in this area. It follows on from PROV’s 2012 Issues Paper on Structured Data, which recommended evaluation of SIARD.

PROV determined that the tools work and the .siard format, based on XML, is acceptable for long-term preservation. PROV will accept transfers of records from relational databases in .siard format (as VEOs).


What is SIARD?

Developed by the Swiss Federal Archives in 2004, SIARD stands for Software Independent Archiving of Relational Databases. It is a set of tools to convert some proprietary databases into standard SQL (non-proprietary), capturing both data and schema (database structure), and storing them in a single .siard format file. 

See Software Independent Archiving of Relational Databases (SIARD) page for further information.


Challenges with long-term preservation of relational databases

Public records, including State Archives, are increasingly stored in relational databases. The overwhelming majority of business systems are underpinned by a relational database to manage and store the data received and created by the system. Structured Query Language (SQL) is the means by which database structure, tables, relationships and the data itself are created/loaded, altered, accessed and maintained.

The consistent structure and interface provided by relational databases and SQL have enabled business system developers to separate the design of their business application from the data source. This has greatly reduced the cost and increased the reliability of information systems. When viewed from a long-term preservation perspective, however, key differences must be addressed. This includes:

  • Later versions of a Relational Database Management system (RDBMS) product may not read databases created under earlier versions without translation.
  • Similarly, a database created under one RDBMS product may not be usable by another RDBMS product without substantial modification. The problem, then, is that the databases created by RDBMS using proprietary variants of SQL, and proprietary data types and processes, cannot be read by other RDBMS products, nor perhaps later versions of the same product.

For an archiving authority like PROV, the incompatibilities between database products and versions creates problems. PROV cannot keep all the different database products to support the databases we receive. This problem becomes greater when considering the need to manage records in such systems for decades to come. One alternative to this is to convert the database into a standard format. This is where SIARD comes in.


What were the results, discoveries and outcomes of the project?

As a result of this project, PROV has approved SIARD as a long-term preservation format and will accept transfers of records from relational databases in .siard format (as VEOs).

The Department of Health and Human Services (DHHS) collaborated in trials of SIARD. This resulted in transfer of the Plan Room Indexing System (PRIS) database. Their work in preparing and transferring the PRIS database to PROV earned DHHS an award at the 2016 Sir Rupert Hamer Awards for ‘the first successful transfer approach to electronic records held in a database in Victoria.’

The final SIARD Research Project Report, delivered in February 2016, included twenty-eight recommendations.

PROV will be implementing the recommendations progressively through 2016 to 2018.