Evaluation of a Large Migration Project

Abstract

The Danish National Archives (DNA) has ingested structurally heterogeneous public digital records since 1973. The year 2004 saw the creation of a new preservation standard into which it was decided to migrate the above mentioned archival holdings. The main objectives of this operation were to save data from technological obsolescence and to reduce the cost of both access and future migrations by streamlining the collection. The project costs approximately 30 FSCs (one ‘FSC’— Format and Structure Conversion—is the way the project’s project management measured 1 person-year, and equals 1,291 person-hours). The total sum of purchasing software, hardware and external services amounted to around 135.000 Euros. The project migrated data from both relational and hierarchical databases (for instance ERDMS and registries), and included the digitisation of audio, video as well as paper documentation. The registries counted for example the first Civil Registration System from 1968 and the State Tax Administration's final equation from 1970. Data and documentation made up a total of about 1.7 TB, consisted of 11,187 files scattered in almost 200 different structures, and constituted more than 2,000 information packages (IPs). The overall technical objective of the migration was defined by the aforementioned preservation standard, which required: • Common format for data files • Common structure of documentation, metadata and documents • Common format for documents (TIFF). The project's main objectives were achieved, since all records were migrated, except the film collection. The goal of making access and future migrations easier was also reached, but a fully automatic migration of the collection is not yet entirely possible. The overall conclusion is that the migration project, which, to our knowledge, is the first of its kind, was of very high quality, both in terms of planning, execution and product. A few main conclusions are: • Standardisation of data is a prerequisite for an economically sound digital preservation: It took about 70 times longer to migrate an older, non-standardised IP, than a newer, standardised one. • Inadequate feasibility studies remind us that the timing in digital preservation saves money: The condition of the magnetic tapes was examined through spot checking 10 years prior to the migration project and found satisfactory, which turned out to be an erroneous, expensive conclusion. • The technical infrastructure suffered from a number of shortcomings and late decisions, resulting in precious loss of time.

Details

Creators

Alex Thirifays; Barbara Dokkedal; Anders Bo Nielsen

Institutions

Date

Keywords

singapore; migration; evaluation; database

Publication Type

paper

License

CC BY-SA 3.0 AT

Download

534994 bytes

View This Publication