This research investigates interoperable query processing with heterogeneous networked information servers. Servers may be databases, image servers, WAIS servers, or Internet search engines. The sources may support OLE (Object Linking and Embedding) DB or ODMG/OQL (Object Data Management Group/Object Query Language)-like interfaces. Architectures are based on mediators (based on some common data model), and wrappers (that provide answers from sources in the common model and format). This research defines WHOL, a High Order Wrapper Language, to extend wrapper technology and support mediator queries that cannot be directly evaluated against the sources. Techniques for evaluating WHOL queries are developed. WHOL supports scalability, as new (dissimilar) sources are incorporated, and extends wrappers into active brokers of information. Techniques for flexible mediator query processing are developed, for wrappers whose capability ranges from simple parameterized functions, to support complex optimization and evaluation engines. Cost models for mediator query processing, and semantics for query evaluation in wrappers with unavailable sources, are also developed. Algorithms are developed for wrapper query processing with time-varying descriptions, and other meta-data, on sources. Solutions will be incorporated within the Disco architecture being developed in INRIA, France. This research will extend the capability and scalability of wrappers and mediators in heterogeneous networked architectures
