The design and experimental evaluation of an information discovery mechanism for networks of autonomous database systems

An approach and mechanism to support the dynamic discovery of information units within a collection of autonomous and heterogeneous database systems is described. The mechanism is based upon a core set of database constructs that characterizes object database systems, along with a set of self-adaptive heuristics employing techniques from machine learning. The approach provides an uniform framework for organizing, indexing, searching, and browsing database information units within an environment of multiple, autonomous, interconnected databases. The feasibility of the approach and mechanism is illustrated using a protein/genetics application environment. Metrics for measuring the performance of the discovery system are presented and the effectiveness of the system is thereby evaluated. Performance tradeoffs are examined and analyzed by experiments performed, employing a simulation model.<>