MSE Wang Institute
My research group investigates infrastructure for distributed digital libraries. Our approach is component-based: we view the digital library infrastructure as a toolbox of service components (repository services, index services, etc), which can be combined (or federated) into digital libraries. This component-based infrastructure is inherently extensible and flexible: components can be evolved independently, new component services can be introduced into the infrastructure, and federations of services can be created and configured according to individual requirements.
This research builds on our design and implementation of Dienst, a protocol and reference implementation for distributed document libraries. Dienst is the technical foundation of the NCSTRL, the Networked Computer Science Technical Research Library, which is administered by our group. NCSTRL is a distributed digital library of computer science research reports from over 100 institutions worldwide.
Currently our infrastructure research is focusing on three areas:
1. We are investigating an architecture, FEDORA, that specifies how content is stored in digital libraries. FEDORA uses a generic abstraction, a digital object, for encapsulating diverse content, making it available in various forms, and providing access control to protect intellectual property.
2. We are studying how to describe and administer collections in digital libraries configured as distributed systems. We view a collection service as an entity that stores and provides metadata about a set of digital objects that conform to a collection criterion.
3. We are investigating techniques to improve distributed searching. In particular, we are interested in how to route queries to replicated and disjoint index servers using metadata provided by those servers.