In order to address the numerous issues that underly the design and development of a successful digital library, we have chosen to base the Alexandria Project on key ideas that will maximize the likelihood for building a scalable system possessing both high-performance and reliability within the four-year life of the project.
The main function of the system will be to allow users to search for, examine, and retrieve spatially-indexed entities from a distributed database either to extract modest amounts of information from appropriate items or to transfer copies of selected items to their own systems, within acceptable response times.
The architecture of the system will involve four relatively simple and independent components whose interactions occur by way of a high-speed wide-area network. These components are a user interface system, a catalogue system, an ingest system, and a storage support system. Combinations of these four components will be permitted at the various distributed sites of the digital library. Much of the Alexandria system architecture is specified primarily in terms of its external interfaces. Note that internal interfaces, such as those to mass storage and DBMS, must be precisely specified. The goal is to make these specifications as vendor-neutral as possible, so that the architecture may take maximum advantage of progress in the marketplace. Furthermore, the system architecture will make maximal use of existing standards. In order to speed the development of a reliable testbed facility, the design and development of the ingest and storage components will be based mostly on the integration of available components, with minimal research intervention.
The design of the user interface system will initially be based on the results of a digital library project already in existence at UCSB. This project has led to a significant understanding of user requirements and various user scenarios. Two experimental systems will be developed: a rapid prototype will be built within the first few months of the project's beginning, using software provided by a private sector partner. This system will be used to test and enhance our knowledge of user requirements. The main testbed system will be developed and tested over the entire four-year period of the project, and will incorporate findings from the rapid prototype system. The initial set of spatially indexed materials that will be supported by the experimental systems will include a number of important and highly-used digitized collections of images, maps and other items. The number of sites that will be initially supported in the distributed network will be relatively small.
In order to accommodate digitized datasets of a variety of types and representational formats, we will adapt an appropriate data model that has been developed for scientific data base systems that typically involve large volumes of heterogeneous, spatially-indexed data. With respect to accessing appropriate subsets of our collections, the research will focus on using a content-based search, with support for such search being accommodated through the use of several levels of information abstraction, from raw pixel-based images and map feature collections to discipline-specific interpretation and automated feature extraction.
The requisite levels of performance that will make the system truly usable will be based on a three-fold strategy that includes: appropriate data models and data structures, including hierarchical structures that are appropriate for browsing and access operations on large entities; appropriate parallel computing support for several computationally-intensive aspects; the linking of components within and between sites with the use of high-speed networks. The system will be scalable over several orders of magnitude of database volume to accommodate future applications that range from school and community libraries through utility companies, local government departments, research libraries, and resource agencies.