Archivematica is a digital preservation system that automates the process of preparing digital objects for ingest into a repository and an access system, ingesting them into archival storage and providing access to the archived material as well as uploading access copies to an access system. The process is monitored and controlled through a Web-based dashboard that co-ordinates a suite of micro-services. It relies on normalisation with preservation as the original object and comprehensive PREMIS metadata in METS.xml as its primary preservation technique.
This project is managed by Artefactual Systems. It began in collaboration with the UNESCO Memory of the World's [http://portal.unesco.org/ci/en/ev.php-URL_ID=1720&URL_DO=DO_TOPIC&URL_SECTION=201.html Subcommittee on Technology, the City of Vancouver Archives, but continues active development along with its partners at the University of British Columbia Library, the Rockefeller Archive Center, Simon Fraser University Archives and Records Management, Yale University Libraries and a number of other collaborators.
Licensing and cost
1.0 (September 2013 / packages December 2014)) is the first production (non-beta) version of the software. The source code for the tool is available from GitHub.
Platform and interoperability
Archivematica may be installed directly on a Linux system, and specifically targets Long Term Support releases of the Xubuntu operating system. Support is included for using Archivematica as a preservation backend for DSpace, a front end for CONTENTdm access, a front end for LOCKSS storage, and a front end for access using AtoM, which comes bundled with the software.
Archivematica uses a micro-services approach, which means it acts as a wrapper for many task-specific applications such as the BagIt library, Clam Anti-Virus, DigiKam, FFmpeg, FITS (File Information Tool Set), ImageMagick, Inkscape, OpenOffice.org, and 7-Zip. The typical workflow is for the curator to assemble a transfer package in the filesystem: a script is provided for setting up the right folder structure or the structure can be assembled manually for some workflows, then digital objects are added to one folder and contextual information (submission documentation in the form of e.g. transfer forms, donation agreements) to another. The package is moved to an input folder 'watched' by the main Archivematica Web tool. Through the Web interface, the curator can decide to accept or reject the transfer. If the transfer is accepted, the tool performs an initial analysis – calculating checksums, assigning UUIDs, scanning for viruses, identifying formats, extracting metadata – and then offers to create a Submission Information Package (SIP); it is also possible to create one or more SIPs manually. Metadata (simple Dublin Core and PREMIS 2.0 rights/restrictions) can then be added to the SIP before it is ingested. At ingest, the curator can choose various routes such as Preservation (where the digital objects are normalised to archival formats and transformed into an Archival Information Package, or AIP), Access (where the digital objects are normalised to dissemination formats and transformed into a Dissemination Information Package, or DIP), repackaging without normalisation, or many combinations of the aforementioned. Further functions are provided for moving AIPs into archival storage and uploading DIPs to the Qubit access portal. Workflows and decision points are configurable via preconfiguration settins in the administration tab of the web-based dashboard.
Documentation and user support
The online documentation for Archivematica includes a User and an Administrative Manual, a walkthrough tutorial, screencasts, a requirements specification (including use cases, activity diagrams, recognised significant properties of various media and media preservation plans) and a description of the technical architecture. Community support is available through the Archivematica Discussion Group. Artefactual Systems, Inc., the primary developer of Archivematica, also offers support options.
The majority of operations are accomplished through a simple Web-based graphical user interface. For some, such as the manual construction of Submission Information Packages, a graphical file management tool (Thunar with additional scripts) is provided, though in future releases such tasks may be supported by the Web interface instead. Reports on the ease of installation and the robustness of the system are mixed but improving; see for example the experiences of Bonnie Weddle and Angela Jordan with version 0.7, and Jenny Mitcham with version 0.9.
The system is easy to use, though as it draws heavily on the OAIS Reference Model some familiarity with that model is needed to understand the workflows Archivematica supports. If using the virtual machine image, some experience of running virtual machines is advantageous but not essential. If installing directly on a Linux desktop or server, a little more technical expertise is required (e.g. for setting up ports correctly). Currently it is only possible to customise the normalisation routes by editing the commands stored in the system's MySQL database.
The functionality of Archivematica is clearly based on that defined by the OAIS Reference Model. The Archival Information Packages generated by the system use the BagIt packaging format, in conjunction with a METS packaging manifest incorporating PREMIS metadata. Metadata entry is through the profile of Dublin Core used by ICA-AtoM (Qubit).
Influence and take-up
Archivematica is used by at least 30 organisations.