Modeling Information System Behavior with Dynamic Relations Nets
Laurent Allain (Institut Superieur d'Electronique du Nord, Departement Informatique, France)
Pascal Yim (Ecole Centrale de Lille, France)
Abstract: In this paper we highlight three main qualities for a processing model: processing abstraction, dynamic behavior and graphical representation. We define a model closely related to high-level Petri Nets. Dynamic Relations Nets (DRN) allow the specification of data, processing, events and constraints within a unique graphical representation. Annotations of the net use a set based abstract language. Constraints arise from three levels: from places (related to the notion of abstract type), from markings (we can then express global constraints between places), and from transitions (in order to specify processing as state transformations). The DRN formalism has been successfully applied to a number of case studies. In this paper, we develop the standard `IFIP case , which has been handled with a lot of modeling methods. A DRN specification has a well defined operational semantics. Therefore a DRN can also be viewed as an executable specification of information systems. We briefly introduce a tool designed to operate an application developed with DRNs, namely NetSpec, based on the use of an active database management system. This tool allows an automated code generation (C/SQL) from a DRN specification.
Keywords: Petri nets, code generation, constraints, dynamic behavior, information systems.
Categories: D.2.2, D.3.4, H.4.2