DIGITAL LIBRARY ARCHITECTURE

In the following paragraphs an architectural approach to the digital library will be developed, which is based on taking the fundamental capabilities, introduced above, as the fundamental requirements the architecture must satisfy. I begin with the following notional architecture for digital libraries:

• A digital library approach to information management depends fundamentally upon a distinction between data and metadata. Metadata provide external classifying and organizing relations for data that may be unstructured, complex, or very large.
• Middleware services such as search, asset protection, and retrieval processes depend on metadata. Since metadata refers to data, which may be stored in separate hierarchical storage subsystems, integrity of reference must be maintained between metadata and data.

Figure 4.1 illustrates these fundamental distinctions between data management and metadata management, and between ingestion and utilization of content.

Operational Architecture

Operational architecture is an information management system represented in terms of the business processes it supports, and how information related to conduct of the business processes passes through the system's components.

The example shown in Figure 4.2 is an enterprise that conducts training by utilizing an extensive computer-based simulation system. The operational (business) processes, most obvious in the example, depend on the timely and well-organized capture of training information as it happens, and both contemporaneous and retrospective search and retrieval of information from a training event. Although the information is generated in several different enterprise domains (eight in the example), effective utilization of information often depends on cross-domain searches and retrievals. Therefore, digital library services must provide information interoperability in middleware.

Fig. 4.1. Notational architecture--building blocks of information to enhance existing functions and enable new operational capabilities.

Fig. 4.2. Operational architecture.

Technical Architecture

A technical architecture breaks down operational (business) processes into functional components and capabilities (Figure 4.3). Hardware and software implementations are still not resolved.

The utilization of digital library materials depends on the existence of metadata to give an efficient and accurate view of content. Metadata must be created as content is added to the digital library. Metadata and data must be bound together logically, and there must be a robust underlying technology to manage the logical connection through time, across platforms, and over geographical separations, all on a networked, distributed system.

Fig. 4.3. Technical architecture.

Systems Architecture

A systems architecture shows the technology enablers and their inter-relationships. In Figure 4.4, the digital library is a centralized subsystem that interacts with a variety of data producers and consumers within a complex distributed system.

Fig. 4.4. Systems architecture.

A fully detailed systems architecture resolves into software and hardware systems. Desirable systems properties such as scalability and extensibility can be taken into account at the systems architecture level. The systems architecture is rationalized relative to the operational and technical architectures.