VACANS-a tool for the validation of CAN-based applications

Abstract

To verify a priori if time constraints of a distributed real time application will be respected, one needs to develop a model suitable for the verification. The paper presents a method for modeling CAN based applications. The models are well suited to performance evaluation for the verification of time constraints. This approach is supported by the tool VACANS, integrating on the one hand a simulation method which allows it to take into account actual needs and to furnish real system performance results and on the other hand the analytic method developed by K. Tindell and A. Burns (1994) which provides the worst case message response time. This tool supports the specification of an application in terms of tasks, inter task data flow task behavior, mapping of tasks onto nodes and CAN network. Such a specification is automatically translated into a simulation model and, when the application only generates periodic messages, translated into an analytic model. The simulation is carried out using the Opnet simulation tool. The user can then verify if the time constraints are met in the actual case or in the worst case. We show the pertinence of the modeling approach and of the VACANS by applying them to an industrial case study.