csm xml
modules RFT xml
GreatSPN GSPN Format
SHARPE FT Format
GreatSPN Format GSPN
GreatSPN GSPN Format
Results Results
Results XML Submodels
ssm xml
modules PFT2PN
Translator
RFT2PN Translator
Figure11: The interfacetoSHARPEandGreatSPN
bydeninga properback-endinterface versus dierentsolvers. The overall approach is
partiallysupportedbythe DrawNET++ framework. Withrespectto otherapproaches,
weallowtoeasilydesignnewformalismsandmodelsstructuringandcompositionschemes,
aswellassupporttointeroperabilityofdierentsolvers.
The application example used throughout the paper concerns the construction and
analysisofdependabilitymodelsusingtwodierentformalisms,namelyFTsandGSPNs.
The proposedmethod however isbynomeans limitedto thisapplicationdomainnor to
theseformalisms,onthe contrarythe exibilityofthe approachallowstoapplyittoany
applicationareaandmodeldescriptionformalismandrelatedsolutionmethodsandtools.
The dependabilityexample considered in the paper is based on a user level
formal-ism appearing as a simple extension to FT, called RFT, includingRepair Blocks (RB);
bothformalismscanbeeasilydenedwithinDrawNET++usingitsMetaformalism
lan-guage. SeveraldierentrepairstrategiescanbedenedandembeddedintotheRBnodes,
in the form of a parametricGSPN model representation, with an interface suitable for
compositionwiththeGSPNtranslationoftherepairablesubtreesintheFT.
Modelscan be structuredin a hierarchyof submodels, allowing to both managethe
complexity of the model design process, and also partition the model in such a way
that it suggests how a multi-solution algorithm may be applied (with the support of
DrawNET++). Anarchitecturehas beendenedforaDrawNET++ back-endthat
con-stitutesaninterfacebetweenDrawNET++andtheSHARPEFTsolverandtheGreatSPN
GSPNsolver. The basic ideaconsists of repeatedlyapplyingthe mostappropriate
solu-tionalgorithm tothe FTsubmodels(starting fromtheinneronesin the hierarchy), and
bringingtheresultbackintotheupperlevelmodels. Anadditionalmodulecouldbe
inte-gratedinthepostprocessor,implementinganalgorithmtochecktheproposedhierarchical
structure,andpossiblymodifyit(inausertransparentway)forachievingamoreeÆcient
combination ofsolutiontechniques.
Inthe proposed example the hierarchical nature of the model suggests a simple
hi-erarchical solution scheme, however any (more complex) scheme might be embedded in
the back end, based on the assumption that the model structure (implicitly or
explic-itly)includesalltheneeded informationto decidehowto decompose itforapplyingthe
multi-solutionalgorithm, andthat thereis awayto make the solvers interact (typically
byfeedingthe resultsfromonesolverintothe inputmodelofanother solver). Ofcourse
theimplementationofthebackendrealizingtheconnectionbetweendierentformalisms
andsolvers isnot trivial,however we believe that thisapproach isappealing because it
enablesthe reuseof existing solvers, andallowsto quickly experimentin a exibleway
combinations ofdierentsolvers.
[1] A. AnandandA.K. Somani. Hierarchical analysisoffault trees withdependencies,
usingdecomposition, InProc. IEEEAnnualReliability and Maintainability
Sympo-sium,pp. 69-75, 1998.
[2] P.Ballarini,S.DonatelliandG.Franceschinis.ParametricStochasticWell-FormedNets
andCompositionalModelling,InProc.21stInternationalConferenceonApplication
and TheoryofPetri Nets,Aarhus,Denmark, June26-30,2000.
[3] BernardiS.,DonatelliS.,HorvathA.,Specialsectiononthepracticaluseofhigh-level
PetriNets: ImplementingCompositionalityforStochasticPetriNets.Int.Journalof
SoftwareTools for Technology Transfer (STTT), SpringerVerlaged.-vol.3issue4,
August2001, pages417{430.
[4] A. Bobbio,G.Franceschinis,R.Gaeta andG.Portinale. DependabilityAssessment
ofanIndustrialProgrammableLogicControllerviaParametricFault-TreeandHigh
LevelPetriNet, InProc.9thInternationalWorkshoponPetriNetsandPerformance
Models,Aachen,Germany, Sept.2001.
[5] G.Booch. ObjectOrientedDesign, BenjaminCummings,RedwoodCity,CA,1991.
[6] R.M.L.R.Carmo,L.R.deCarvalho,E.deSouzaeSilva,M.C.DinizandR.R.Muntz.
Performance/availabilitymodelingwiththeTangram-II modelingenvironment,
Per-formance Evaluation33(1998),pp.45-46.
[7] G.Chiola,G.Franceschinis,R.Gaeta, andM. Ribaudo. GreatSPN1.7: GRaphical
EditorandAnalyzer forTimedandStochasticPetriNets, PerformanceEvaluation,
vol.24,no.1-2,Nov.1995,1995, pp.47-68. http://www.di.unito.it/greatspn/
[8] G. Chiola, C.Dutheillet, G. Franceschinis, and S. Haddad. Stochasticwell-formed
colourednetsforsymmetricmodellingapplications, IEEETOC,vol.42pp.1343-1360,
1993.
[9] G. Clark, T. Courtney, D. Daly, D. Deavours, S. Derisavi, J. Doyle, W. Sanders,
P. Webster. The Mobiusmodeling tool, InProc. 9th Int. Workshop on Petri Nets
and PerformanceModels,Aachen,Germany,Sept.2001,pp.241-250.
[10] D.DeavoursandW.Sanders. Mobius: FrameworkandAtomicModels, InProc. 9th
International Workshop onPetri Netsand Performance Models, Aachen, Germany,
September2001.
[11] R. Gaeta, A. Bobbio, G. Franceschinisand G. Portinale. Exploiting Petri Nets to
supportFaultTreebaseddependabilityanalysis,InProc.8thInternationalWorkshop
onPetri NetsandPerformanceModels, Zaragoza,Spain,September 1999.
[12] Y.DutuitandA.Rauzy.Alinear-timealgorithmtondmodulesoffaulttrees,IEEE
TransactionsonReliability,Vol.45,No. 3,September1996.
[13] G. Franceschinis, M. Gribaudo, M. Iacono, N. Mazzocca and V. Vittorini.
DrawNET++: Model Objects to Support Performance Analysis and Simulation
of Complex Systems, In Proc. 12th International Conference on Modelling Tools
(TOOLS 2002), London, UK, Lecture Notes in Computer Science, Volume 2324,
Springer,pp.233-238,April2002.
[14] M. Gribaudo, A. Valente. Framework forGraph-based Formalisms, InProceeding
ofthe rstInternational ConferenceonSoftwareEngineeringApplied toNetworking
and Parallel Distributed Computing 2000, SNPD'00, pages 233-236 Reims, France,
May2000, ACIS.
[15] H.HenleyandH.Kukamoto.ReliabilityEngineeringandRiskAssessment,
Prentice-Hall,EnglewoodClis,1981.
[16] N.Leveson. Safeware: SystemSafetyandComputers, Addison-Wesley,1995.
[17] K. Jensen. Coloured Petri Nets - Basic Concepts, Analysis Methods andPractical
Use,Vol.1: BasicConcepts, EATCSMonographsonTheoreticalComputerScience.
Berlin: Springer-Verlag,1992.
[18] R.A. Sahner, K.S. Trivedi, and A. Puliato. Performance and Reliability
Anal-ysis of Computer Systems, An Example-Based Approach Using the SHARPE
Software Package, Kluwer Academic Publisher, November 1995, Boston, 1995.
http://sun195.iit.unict.it/GISharpe/index.html