Titel (eng): Studies on the scalability of web preservation: Paper - iPres 2013 - Lisbon

Autor: Blevins, Rory
Patel, Ismail
O'Sullivan, Jack
Hunter, Ashley
Sharpe, Robert
Sinclair, Pauline

Beschreibung (eng): This paper describes a mechanism for improving the scalability of preservation actions on large linked archives, such as WARC and ARC files produced from the archiving of web sites.
To enable accurate but efficient preservation actions, information on the files embedded within a container object, such as the file formats of the embedded files, are aggregated and recorded as properties of the container object. This occurs during the ingest of objects into the archiving system, specifically at the characterization stage when files are identified and validated. To ensure that the details of all embedded files are also recorded, nested archives are recursively unpacked and their contents characterized to identify all files in a package. Information about the embedded files is then stored as properties of the container object: this allows us to efficiently aggregate information about the contents of a container as queryable properties of the container.
This storage of the embedded file type information on the container object reduces the number of objects and properties which have to be queried to perform a preservation action, such as migration to a more recent file type. The database can be queried for a specific file type, and all files of that type, and archives containing files of that type will be returned without needing to query each embedded object individually.
Archives containing files in need of preservation are temporarily unpacked and the files in need of transformation identified and migrated. Following the preservation action, the internal links within the archive are updated to maintain the integrity of the archive and the modified objects are re-ingested back into the system.
This approach results in minimal extra overhead at the ingest stage of preservation, but substantially reduces the number of entities which need to be queried to identify objects at risk when
performing preservation actions. In the case of large web archives, this may be several orders of magnitude, producing a corresponding increase in performance and scalability.

Sprache des Objekts: Englisch

Rechte: Creative Commons Lizenzvertrag
Dieses Werk bzw. dieser Inhalt steht unter einer CC BY-SA 2.0 AT - Creative Commons Namensnennung - Weitergabe unter gleichen Bedingungen 2.0 Österreich Lizenz.

CC BY-SA 2.0 AT

http://creativecommons.org/licenses/by-sa/2.0/at/

Klassifikation: Scalability, Web Archiving, Characterization, iPRES, Lisbon

Conferences, Conference 2013

Besuche auf dieser Seite:

Downloads:

Mitglied in der/den Collection(s) (3):
o:424738 Openaire v3.0 collection
o:378098 iPRES 2013 - Proceedings of the 10th International Conference on Preservation of Digital Objects: iPRES 2013 - Lisbon
o:168770 Open Access Documents in Phaidra
Permanent Identifier