A model-based fault-tolerant CSCW architecture. Application to biomedical signals visualization and processing

The paper describes a methodological approach that uses Petri nets (PNs) and Time Petri nets (TPNs) for modeling, analysis and behavior control of fault-tolerant computer supported synchronous cooperative work (CSSCW) architectures inside which a high level of interactivity between users is required. Modeling allows architectures to be formally studied under different functioning conditions (normal communications and deficient communications). Results show that the model is able to predict interlocking and state inconsistencies in the presence of errors. TPNs are used to extend PN models in order to detect communication errors and avoid subsequent dysfunctions. The approach is illustrated through the improvement of a recently presented collaborative application dedicated to biomedical signal visualization and analysis.