Transcript Preservation of digital records in the long term

Approaches to database archiving
at the Danish National Archives
Open Planets Foundation Hackathon 2012-02-07 Database Archiving Event
Databases in any flavour
• 1976 – mostly hierarchical databases or network
databases
– system independent archiving, but very flexible requirements
– fixed or variable field length
– any kind of data type
- alphanumeric, packed decimal, binary, floating point
– any kind of character set
- BCD, EBCDIC, ASCII, and many proprietory variants of these
Open Planets Foundation Hackathon 2012-02-07 Database Archiving Event
Relational databases only
• 1998, 2000, 2004 – all databases must be submitted as
relational databases
– system independent archiving, strict requirements
– hierarchical databases must be migrated to relational
databases
– markup language used to describe stucture
– fixed or variable field length
– data types limited to the most common ISO data types
– character set limited to ISO 8859-1
Open Planets Foundation Hackathon 2012-02-07 Database Archiving Event
The search for the next rdb archving
format
• 2005 – still searching for a more standardarised (closer
to SQL) and widespread rdb archiving format
• Options:
–
–
–
–
–
further developing our own format
ADDML
SIARD
DBXML
and a few other even smaller projects
• 2007 First International Workshop on Database
Preservation (PresDB’07)
Open Planets Foundation Hackathon 2012-02-07 Database Archiving Event
SIARD chosen
• SIARD = Software-Independent Archiving of Relational
Databases
– XML markup of SQL DDL (SQL:1999)
– XML markup of data
• Developed by the Swiss Federal Archive
• Chosen as archive (preservation) format for databases in
the European PLANETS project
Open Planets Foundation Hackathon 2012-02-07 Database Archiving Event