Digital Preservation Case Study: National Archives and Records Administration

1
Digital Preservation Case Study National
Archives and Records Administration

• Paul Watry
• University of Liverpool

2
NARA prototype

• World's largest digital preservation initiative
• Runs until 2012 with 350m spend.
• Deals with all aspects of information life-cycle
  management
• Looks at any data, not just HTML or web-based
  data.

3
Major issues

• Preservation of records is a two-fold problem
• First, the preservation of the original record
  and making it available over time
• Second, creating a digital representation of the
  original record in a software independent fashion
  and using this proxy for the original record.
• The technologies required to solve these twin
  challenges come from the data grid, persistent
  archive, and digital library communities.

4
Combination of technologies

• Data Grids provide the means to manage
  information (the actual bits that comprise
  documents).
• We are talking about petabytes of data.
• Digital Library technologies provide ways of
  accessing data (finding the documents you
  want).
• Persistent Archives technologies provide ways of
  presenting (or viewing) and manipulating
  documents.
• NOT simply the web (email, scientific simulation
  programs, etc.)

5
Joint Work Liverpool and SDSC

• SDSC is providing the data grid technologies
  through the SRB (pronounced serb).
• Liverpool is providing the digital library and
  persistent archive technologies (Cheshire and
  Multivalent systems).
• San Diego provides the capability of storing and
  keeping track of the data (including replication)
  over distributed supercomputers
• Liverpool provides the capability of finding,
  presenting, and reusing the data held in these
  large repositories.

6
Establishing Provenance or Chain of Custody

• Automating archival processes
• Ingest into the data grid. Who created data? What
  format? Who has access to view or edit data?
• Developing collection management systems
• How do you keep track of the data in distributed
  data grids?
• Developing digital ontologies
• Once you have stored the data, how do you
  retrieve it? You will need to view it where there
  is no software? How do you reuse it?

7
Obsolescence of digital document data formats

• Big problem how do you view documents for which
  there is no software.
• Strategies in the past migration, emulation,
  universal format (PDF).
• Problems
• Migration degrades data over time (20 years)
• Emulation you have to keep a room full of
  functioning machines.
• Universal format get a life

8
Multivalent preservation architecture

• A browser which will read documents from the
  original bitstream through the use of media
  adapters.
• Does this by transforming the document into
  browser's internal structure.
• We can then present and manipulate the document
  independently of the software or infrastructure
  that created it.
• The manipulation capability means that we can
  reuse legacy documents in any number of ways.