DrawNet++: a Flexible Framework for Building Dependability Models
G. Franceschinis, M. Gribaudo, M. Iacono, V. Vittorini, C. Bertoncello
y1. DrawNet++ quick overview and context
The DrawNet++ project addresses the compositional construction of dependability models [3, 4]. Its main goals are to provide: a) a GUI to any graph-based formalism; b) a support to the design process of dependability models, according to concepts inspired by object orientation (OO);
c) a user friendly front-end for different classes of analy- sis/simulation tools.
Such features enforce the rapid prototyping and the reuse of sub-models. The formalisms adopted to implement the (sub)models may differ, by allowing to study both compo- sitional problems related to the usage of a single specifica- tion formalism and interoperability issues within a multi- formalism environment. A prototype post-processor has been implemented to interface a few analysis/simulation tools and retrieve/show performance results. By now, Para- metric Fault Tree (PFT) [1] and Stochastic Well Formed Nets (SWN)[2], are supported. An XML description of the user defined models is generated by the tool and XSL styles are used to translate the XML representation into a syntax suitable for the specific analysis tool. We use the DrawNet++ framework to build compositional SWN mod- els and models of highly redundant systems by combin- ing PFT with SWN [1]: since state space analysis tools are needed to represent dependencies between components, repair, transient faults etc, we model them by SWN sub- models and compose such predefined blocks with a SWN model of the system obtained by automatically translating a PFT model. In this context we use the DrawNet++ a) to define and draw composed models (e.g. PFT and SWN blocks); b) to produce the syntax representation of the entire model and c) to present the results computed by the solvers.
2. DrawNet++ demonstration
The goal of this demonstration is to present (1) the defini- tion of a graph based formalism, including compositionality
G.Franceschinis and C.Bertoncello are with the Univ. del Piemonte Orientale, Alessandria, Italy; M.Gribaudo is with the Univ. di Torino, Torino, Italy; M. Iacono is with the Seconda Univ. di Napoli, Aversa, Italy; V.Vittorini is with the Univ. di Napoli ”Federico II”, Napoli, Italy.
yThis work is partially supported by the MIUR (Project ”ISIDE”).
issues; (2) some guidelines for a dependability model con- struction cycle from PFTs to SWNs; (3) the OO capabilities that can be exploited in the model design process.
The first point addresses the XML specification of a graph based formalism and the automatic creation of a GUI to edit and manipulate (sub)models (Fig.1). The second point deals with the usage of XSL style sheets and the post- processing of the models in order to operate the conversion from a PFT model into a SWN model; this will be demon- strated by means of a case study [1]. Finally, the third point is aimed to show the mechanisms to define classes and ob- jects that represent sets of system component models, create hierarchies, implement information hiding capabilities, and specify template objects.
Figure 1. A PFT example.
References
[1] A. Bobbio, G. Franceschinis, L. Portinale, and R. Gaeta. De- pendability Assessment of an Industrial Programmable Logic Controller via Parametric FaultTree and High Level Petri Net.
Proc. of PNPM01, Aachen, Germany, September 2001.
[2] G. Chiola, C. Dutheillet, G. Franceschinis, and S. Haddad.
Stochastic wellformed coloured nets for symmetric modelling applications. IEEE TOC, 42:1343–1360, 1993.
[3] G. Franceschinis, M. Gribaudo, M. Iacono, N. Mazzocca, V. Vittorini. DrawNET++: model objects to support perfor- mance analysis and simulation of complex systems. Proc. of TOOLS2002, London, UK, April 2002.
[4] M. Gribaudo, A. Valente. Framework for Graph-based For- malisms. Proc. of SNPD’00, Reims, France, May 2000.