Active database

Abstract: A crucial requirement for active databases is the ability to analyse the behaviour of the active rules. A particularly important type of analysis is termination analysis. We define a framework for modelling the execution of active rules, based on abstract interpretation. Specific methods for termination analysis are modelled as specific approximations within the framework. The correctness of a method can be established by proving two generic requirements provided by the framework. This affords the opportunity to compare and verify existing methods for termination analysis of active rules, and also to develop new ones.

Abstract: The association is a native concept from relational databases, one that has been adapted to object oriented (OO) modelling. It is an interesting operator used to describe links among objects of a system, commonly included in the most popular diagram-based OO methodologies. However, those methodologies sometimes present a lack of formality that may undermine its use. In this paper we formalize the semantics of associations. Firstly, we will describe an OO model based on different kinds of constraints. Some of them will be especially useful for describing the semantics of associations. Finally, we will present some remarks about implementation by means of triggers, a new feature incorporated in databases to specify an inner active behavior.

Abstract: EAR-Impish is a database management system (DBMS) that extends the entity-relationship (E-R) concepts and provides its users with several powerful capabilities. It supports integrity constraints representation and data consistency checking. It allows management of time-related concepts and versions of entity and relationship instances. Because it can automatically insert or delete instances, and because it supports definition of demons, EAR-Impish can be considered as an active database system. EAR-Impish also allows the definition and manipulation of hypotheses, a feature that is useful in decision help systems. Finally, it supports documents, which are large attributes, and a simple consistency control of document contents. EAR-Impish characteristics have been specified to fulfil the need for software engineering environments to have adapted and powerful information systems. EAR-Impish fully implements basic E-R concepts, and can be used as an E-R DBMS. However, as it is extended with new concepts, has deductive and automated capabilities, and is coupled with Prolog, EAR-Impish is open and useful to both traditional applications and expert systems. EAR-Impish combines database, file, and logical systems. It implements E-R concepts on top of an extended relational database system, Impish.

Abstract: Databases nowadays used in all kinds of application areas often differ greatly in a number of properties. These varieties add complexity to the handling of databases, especially when two or more different databases are dependent. The approach described here to propagate updates in an application scenario with heterogeneous, dependent databases is the use of a rule-based mediator. The system EruS (ECA rules updating SCOP) applies active database technologies in a bioinformatics scenario. Reactive behaviour based on rules is used for databases holding protein structures. The inherent heterogeneities of the Structural Classification of Proteins (SCOP) database and the Protein Data Bank (PDB) cause inconsistencies in the SCOP data derived from PDB. This complicates research on protein structures. EruS solves this problem by establishing rule-based interaction between the two databases. The system is built on the rule engine ruleCore with Event-Condition-Action rules to process PDB updates. It is complemented with wrappers accessing the databases to generate the events, which are executed as actions. The resulting system processes deletes and modifications of existing PDB entries and updates SCOP flatfiles with the relevant information. This is the first step in the development of EruS, which is to be extended in future work. The project improves bioinformatics research by providing easy access to up-to-date information from PDB to SCOP users. The system can also be considered as a model for rule-based mediators in other application areas.