Versions Compared


  • This line was added.
  • This line was removed.
  • Formatting was changed.


The wide adoption of digital repository software by cultural heritage organizations has led to significant increases in effective management and access to digital assets. While these software systems may provide some digital preservation features, the digital materials and their associated metadata managed by such systems need to be preserved in a way that ensures they will persist over time. This project addresses this need by developing a specification for an integration model that will allow libraries and archives to seamlessly deposit system content into distributed digital preservation systems (DDPs) such as Chronopolis, APTrust, and LOCKSS. This project is funded by the Mellon Foundation. 


While digital repository systems are generally backed-up, this back-up tends to involve managing just one copy of the data. As such, it does not fulfill standards-based digital  preservation requirements, which mitigate risks associated with the lack of geographical redundancy, technological diversity, and also human error or malfeasance, all of which can be accomplished by distributing multiple copies of deposited data across a network of nodes which are based in different geographical locations and managed by different entities using different storage infrastructure stacks. Distributed digital preservation systems (DDPs) like Chronopolis, APTrust, and LOCKSS are designed to fill this preservation gap, and many organizations have adopted these services in conjunction with their local repositories. However, DDP services and local repositories are not designed to work together and their incongruencies present a multitude of issues for organizations wishing to employ both types of systems. Typically, local repositories are driven by access concerns, and their design and internal workflows are directed thusly. For example, these systems tend not to track preservation events (e.g. synchronization timestamps, checksum comparisons) but are more focused on access information (e.g. access counts and downloads). In contrast, DDP services are focused on the status of data among their nodes, so things like synchronization timestamps and checksum comparisons are crucial. Currently, there are no good options for users who are running these systems to efficiently (or automatically) link these two kinds of systems in a way that allows for management of data between them. To address this problem a new specification is needed that will facilitate preservation-quality deposit from local repositories to DDPs.