Constraint and data fusion in a distributed information system

Abstract

Constraints are commonly used to maintain data integrity and consistency in databases. This ability to store constraint knowledge, however, can also be viewed as an attachment of instructions on how a data object should be used. In other words, data objects are annotated with declarative knowledge which can be transformed and processed.

This abstract describes our work in Aberdeen on the fusion of knowledge in the form of integrity constraints in a distributed environment. We are particularly interested in the use of constraint logic programming techniques with off-the-shelf constraint solvers and distributed database queries. The objective is to construct an information system that solves application problems by combining declarative knowledge attached to data objects in a distributed environment. Unlike a conventional distributed database system where only database queries and data objects are transported, we also ship the constraints which are attached to the data. This is analogous to the use of footnotes and remarks in a product catalogue describing the restrictions on the use of specific components.

This research is part of the KRAFT research project [1] funded by EPSRC and BT.

Kit-ying Hui is supported by a grant from BT.