Preservation Environments

The long-term preservation of digital entities requires mechanisms to manage the authenticity of massive data collections that are written to archival storage systems. Preservation environments impose authenticity constraints and manage the evolution of the storage system technology by building infrastructure independent solutions. This seeming paradox, the need for large archives, while avoiding dependence upon vendor specific solutions, is resolved through use of data grid technology. Data grids provide the storage repository abstractions that make it possible to migrate collections between vendor specific products, while ensuring the authenticity of the archived data. Data grids provide the software infrastructure that interfaces vendor-specific storage archives to preservation environments. 1. Introduction A preservation environment manages both archival content (the digital entities that are being archived), and archival context (the metadata that are used to assert authenticity) [8]. Preservation environments integrate data storage repositories with information repositories, and provide mechanisms to maintain consistency between the context and content. Preservation systems rely upon software systems to manage and interpret the data bits. Traditionally, a digital entity is retrieved from an archival storage system, structures within the digital entity are interpreted by an application that issues operating system I/O calls to read the bits, and semantic labels that assign meaning to the structures are organized in a database. This process requires multiple levels of software, from the archival storage system software, to the operating system on which the archive software is executed, to the application that interprets and displays the digital entity, to the database that manages the descriptive context. A preservation environment assumes that each level of the software hierarchy used to manage data and metadata will change over time, and provides mechanisms to manage the technology evolution.