Designing a Bit Preservation System

Abstract

Despite successful operation of DSpace and Fedora repositories, much of the digital preservation work performed at the University of Maryland (UMD) Libraries is focused more on backup and restoration of files than on actual archiving and preservation. The Digital Systems and Stewardship (DSS) division has begun a project to examine gaps, risks, and inefficiencies in the current systems, as well as to build support for new functionalities, such as the curation of active research datasets that are used and can be updated over time. Our first phase of repository development narrows the scope to bit-level preservation and access services. We define bit-level preservation as preservation regardless of the semantics or intellectual content of what is included in a file (content, virus, encryption, intellectual content, etc.), the file format, and how files are semantically interrelated.

During high-level requirements development, it became clear that this bit-level preservation system, rather than representing just a back-end for our existing systems, can provide digital content management services in all phases of content’s lifecycle, including selection, creation, acquisition, and disposition. This paper will discuss the decision to design a bit-level preservation system, our high-level requirements, and our investigation and implementation plans.