Upload
buikhue
View
214
Download
0
Embed Size (px)
Citation preview
dependability engineering & Petri nets October 2014
Z:\Documents\teaching\nl\nl_skript_fm\nl01_framework.sld.fm 1 - 1 / 27
Brandenburg Universityof Technology,
Computer Science Institute
PETRI NET BASED
DEPENDABILITY ENGINEERING
OF REACTIVE SYSTEMS
MONIKA HEINER
[email protected]://www.informatik.tu-cottbus.de
dependability engineering & Petri nets October 2014
[email protected] 1 - 2 / 27
BASIC STRUCTURE
OF REACTIVE SYSTEMS
controller
environment model
plant / devices
actuatorssensors
preprocess
postprocess
dependability engineering & Petri nets October 2014
[email protected] 1 - 3 / 27
feed
bel
t (be
lt 1
)
depo
sit b
elt (
belt
2)
elev
atin
g ro
tary
tabl
e
robo
t
arm
1
arm
2
pres
s
trav
elli
ng c
rane
EX
AM
PL
E, P
RO
DU
CT
ION
CE
LL:
14 s
enso
rs34
com
man
ds
dependability engineering & Petri nets October 2014
[email protected] 1 - 4 / 27
EXAMPLE,CONCURRENT PUSHERS
Pos. 3 Pos. 2
Pus
her
1
R 2
R 1
Pusher 2
M
M
Piece, Pos. 1
Controller 2
R 2
R 1
Controller 1
dependability engineering & Petri nets October 2014
[email protected] 1 - 5 / 27
EXAMPLE,CRUISE CONTROL
❑ pressing on, while car ignition is switched on -> current speed is recorded and
system is enabled
❑ pressing brake, accelerator or off -> system is disabled
❑ pressing resume-> re-enables the system
dependability engineering & Petri nets October 2014
[email protected] 1 - 6 / 27
MOTIVATION
CONCURRENCY IS WIDESPREAD, BUT ERROR PRONE
❑ Therac-25 computerized radiation therapy machine
-> concurrent programming errors contributed toaccidents causing deaths and serious injuries
❑ Mars Rover
-> problems with interaction between concurrenttasks caused periodic software resets reducing availability for exploration
❑ . . .
OBVIOUS QUESTIONS
❑ is a system safe ?
❑ is a system reliable ?
❑ would testing be sufficient to discover all errors ?
dependability engineering & Petri nets October 2014
[email protected] 1 - 7 / 27
PRELIMINARIES
❑ DEPENDABILITY
ability of a systemto fulfill its predefined task(in spite of any hardware and/or software faults)
❑ dependability modelling
❑ engineer’s basic principle:
KEEP EVERYTHING AS SIMPLE AS POSSIBLE!
Which kind of models?
Where do the models come from?
dedicated models for different kinds of properties;
dependability engineering & Petri nets October 2014
[email protected] 1 - 8 / 27
METHODS
MANUAL
COMPUTER-AIDED
FAULT AVOIDANCE
FAULT REMOVAL
FAULT TOLERANCE
FAULT MASKING
FAULT RECOVERY
FAULT PREVENTION
development
operation
DEFENSIVE
DIVERSITY
SOFTWARE DEPENDABILITY
phase
phase
animation / simulation / testing
context checking (static analysis)
consistency checking (verification)
VALIDATION
dependability engineering & Petri nets October 2014
[email protected] 1 - 9 / 27
MODEL BASED
SYSTEM VALIDATION,GENERAL PRINCIPLE
PetrinetzmodelProblemsystem
systemproperties
modelproperties
What was in the beginning ?
dependability engineering & Petri nets October 2014
[email protected] 1 - 10 / 27
MODEL-BASED
SYSTEM VALIDATION,TWO APPROACHES
Petrinetz
Problemproblem
model
implementation
validation
modelling
validation
Petrinetzmodel Petrinetzsystem
Petrinetzsystem
dependability engineering & Petri nets October 2014
[email protected] 1 - 11 / 27
A POSTERIORI MODEL BASED
SYSTEM VALIDATION,PROCESS AND TOOLS
requirements
controller environmentsafety
requirements
compiler modelling
temporal
library
control model
environmentmodel
set oftemporal
composition
systemmodel
verification methods errors /
formulae
logic
functional
inconsistencies
dependability engineering & Petri nets October 2014
[email protected] 1 - 12 / 27
OBJECTTIVE -REUSE OF
CERTIFIED COMPONENTS
REALPROGRAM
DREAMPROGRAM
SAFETY REQUIREMENTS
FUNCTIONALREQUIREMENTS
dependability engineering & Petri nets October 2014
[email protected] 1 - 13 / 27
ANOTHER APPLICATION:BIOCHEMICAL SYSTEMS,
EXAMPLES
❑ metabolic pathways / networks
->stoichiometric relations known
->concentrations of metabolites often known
❑ signal transduction pathways / networks
->stoichiometric relations unknown
->read arcs / test arcs
->inhibitor arcs
❑ gene regulatory networks
-> stoichiometric relations unknown
->mRNA concentrations often known
->protein concentrations are hard to be measured
->often a mixture of metabolic and signal transduction pathways
=>> networks of elementary actions
dependability engineering & Petri nets October 2014
[email protected] 1 - 14 / 27
MODEL- BASED
SYSTEM ENGINEERING
Petrinetzmodel
Problemsystem systemproperties
modelproperties
biochemicalsystem
known
unknown properties
properties
behaviour predictionU
ND
ER
STA
ND
ING validation
GENERALIZATION TO BIOCHEMICAL SYSTEMS
dependability engineering & Petri nets October 2014
[email protected] 1 - 15 / 27
BIONETWORK, EX 1G-PP PATHWAYS
Ru5P4
5
Xu5P
R5P
6S7P
GAP
7E4P
F6P
8
GAP
15
NAD+
+Pi
NADH
G6P F6P10
ATP ADP
FBP11 12
DHAP
13
14ATP ADP
9Gluc
1,3-BPG
ATP ADP
16
ATP ADP
19
NAD+ NADH
203PG
172PGPEP
18PyrLac
NADP+ 2
NADPH 2
2 GSH 2
2 31
2 GSSG
GLYCOLYSIS / PENTOSE PHOSPHATE PATHWAYS IN ERYTHROCYTES
[Reddy 1996]
dependability engineering & Petri nets October 2014
[email protected] 1 - 16 / 27
BIONETWORK, EX 1AS PETRI NET,
VERSION 1
2
22 2
F6PGAP
ATP ADP ADPATP
NADHNAD+ ATP ADP
Lac Pyr PEP 2PG 3PG 1,3-BPG
DHAP
FBP
G6P
Gluc F6P
E4P
GAP
S7P
R5P
Xu5PRu5P
GSH
GSSG NADPH
NADP+
NAD+
ADPATP
Pi
glucose1.spped
NADH
1 2 3
4
5
6 7 8
9 10 11 1213
14
15
1617181920
dependability engineering & Petri nets October 2014
[email protected] 1 - 17 / 27
BIONETWORK, EX2,APOPTOSIS
http://www.genomicObject.net
APOPTOSIS IN MAMMALIAN CELLS
dependability engineering & Petri nets October 2014
[email protected] 1 - 18 / 27
BIONETWORK, EX2,AS PETRI NET,
VERSION 1
s6
s4
s3
s13
s12
s2
s11
s10
s5
s8
s9
s7
s1
Mitochondrion
DNA-Fragment
DNA
DFF40-Oligomer
CleavedDFF45
DFF
Caspase-3
Procaspase-3
Caspase-9
Procaspase-9
(m22)
(m20)
Apaf-1
dATP/ATP
CytochromeC
Bcl-2_Bcl-xL
Apoptotic_Stimuli
Bax_Bad_Bim
BidC-TerminalBid
Caspase-8
Procaspase-8FADD
Fas-Ligand
fas1.spped
dependability engineering & Petri nets October 2014
[email protected] 1 - 19 / 27
BIONETWORK, EX3,POTATO TUBER
R1. SuSy: sucrose synthase Suc + UDP <--> UDPglc + Frc
R2. UGPase: UDPglucose pyrophosphorylaseUDPglc + PP <--> G1P + UTP
R3. PGM: phosphoglucomutase G6P <--> G1P
R4. FK: fructokinase Frc + ATP --> F6P + ADP
R5. PGI: phosphoglucose isomerase G6P <--> F6P
R6. HK: hexokinase Glc + ATP --> G6P +ADP
R7. Inv: invertase Suc --> Glc + Frc
R8. Glyc(b): glycolysis F6P + 29 ADP + 28 Pi --> 29 ATP
R9. SPS: sucrose phospahate synthase F6P + UDPglc <--> S6P + UDP
R10. SPP: sucrose phosphate phosphataseS6P --> Suc + Pi
R11. NDPkin: NDP kinaseUDP + ATP <--> UTP +ADP
R12. SucTrans: sucrose transportereSuc --> Suc
R13. ATPcons(b): ATP consumptionATP --> ADP + Pi
R14. StaSy(b): starch synthesis G6P + ATP --> starch + ADP + PP
R15. AdK: adenylate kinase ATP + AMP <--> 2 ADP
R16: PPase: pyrophosphatase PP --> 2 Pi
dependability engineering & Petri nets October 2014
[email protected] 1 - 20 / 27
BIONETWORK, EX3,AS PETRI NET
22
2
28 29
29
PPase
rStarch
ATPcons(b)
Glyc(b)
geSuc
StaSy(b)
SPP
FKHK
Inv
SucTrans
29 ADPAMP
ATP
28 Pi
28Pi
starch
28Pi
ATP
29 ADP
PP
UDP
29 ADP
ATP
PP
29ADP
ATP
28 Pi
S6P
29 ADP
ATP
29 ADP
ATP
UTP
UDP
G1P
F6P
G6P
UDPglcFrcGlc
eSuc
Suc
7. AdK
6. UGPase
5. NDPkin 4. PGM
3. PGI
2. SPS
1. SuSy
potato.spped
dependability engineering & Petri nets October 2014
[email protected] 1 - 21 / 27
WHY PETRI NETS?
❑ a suitable intermediate representation for
different (specification/programming) languages,
different phases of software development cycle,
different validation methods;
❑ modelling power
partial order (true concurrency) semantics
applicable on any abstraction level
specification of limited resources possible
❑ analyzing power
not restricted to reachability graph
❑ BUT: modelling power <-> analyzing power
❑ integration of qualitative and quantitative analyses
dependability engineering & Petri nets October 2014
[email protected] 1 - 22 / 27
INTEGRATION OF
QUALITATIVE &QUANTITATIVE ANALYSES
net-basedtesting andmonitoring
net-basedqualitativeanalysis
net-basedquantitative
analysis
dependability engineering & Petri nets October 2014
[email protected] 1 - 23 / 27
MODEL
CLASSES
context checking byPetri net theory
verification bytemporal logics
performanceprediction
reliabilityprediction
PETRI NETS
PLACE/TRANSITION
(COLOURED PN)
TIME-DEPENDENT PN
NON-STOCHASTIC
STOCHASTIC
PETRI NET
PETRI NET
PETRI NET
worst-caseevaluation
dependability engineering & Petri nets October 2014
[email protected] 1 - 24 / 27
TOOL
OVERVIEW
❑ Snoopy
design / animation / simulation of Petri nets, e.g.
QPN - XPN - SPN - XSPN - CPN - HPN,
and the coloured counterparts,
... and many more ...
special features
logical places / transitions
macro transition / places
❑ Charlie
standard Petri net analysis techniques, e.g.
structural properties
P/T-invariants
Siphon/Trap Property, rank theorem
reachability/coverability graph
(explicit) CTL model checking
❑ Marcie
QPN - symbolic CTL model checking
SPN - symbolic CSL model checking,
XSPN - simulative PLTLc model checking
dependability engineering & Petri nets October 2014
[email protected] 1 - 25 / 27
CASE STUDIES
ACADEMIC:
❑ botanical garden
❑ low-level mutex algorithm
❑ Dijkstra’s philosophers
❑ Milner’s scheduler
❑ solitaire
❑ . . .
MORE REALISTIC
❑ production cell
❑ concurrent pushers
❑ cruise control
❑ . . .
dependability engineering & Petri nets October 2014
[email protected] 1 - 26 / 27
REFERENCES I
Snoopy
G Czichy (1993)Design and Implementation of a graphical editor for hierarchical Petri net models (inGerman);TU Dresden, Dep. of CS, Master Thesis 1993.
R TIEDEMANN (1997)PED - Hierarchical Petri Net Editor, Manual (in German);BTU Cottbus, Dep. of CS, Internal Techn. Report, May 1997.
T MENZEL (1996)Design and Implementation of a Petri Net Tool Kit Framework Integrating Animation andSimulation (in German); BTU Cottbus, Dep. of CS, Major Individual Project, 1996.
M Fieber (2004)Design and Implementation of a Generic and Adaptive Graph Tool (in German), BTU Cottbus, Dep. of CS, Master Thesis, July 2004
M Heiner, R Richter, M Schwarick (2008)Snoopy - A Tool to Design and Animate/Simulate Graph-Based Formalisms;Proc. PNTAP 2008, associated to SIMUTools 2008, ACM digital library, 2008.
C Rohr, W Marwan, M Heiner (2010)Snoopy - a unifying Petri net framework to investigate biomolecular networks;Bioinformatics 26(7):974-975, 2010.
M Heiner, M Herajy, F Liu, C Rohr, M Schwarick (2012)Snoopy – a unifying Petri net tool;Proc. PETRI NETS 2012, Hamburg, Springer, LNCS 7347, 398-407, June 2012.
Fei Liu (2012)Colored Petri Nets for Systems Biology;PhD thesis, BTU Cottbus, Dep. of CS, January 2012.
M Herajy (2013)Computational Steering of Multi-Scale Biochemical Networks;PhD thesis, BTU Cottbus, Dep. of CS, January 2013.
S Laarz (2013)Scalable Petri nets in Snoopy (in German), BTU Cottbus, Dep. of CS, Master Thesis, Februar 2013
dependability engineering & Petri nets October 2014
[email protected] 1 - 27 / 27
REFERENCES II
Charlie
PH Starke, S Roch (1997) INA - Integrated Net Analyser version 1.7; Technical report, Humbold-Universität zu Berlin, 1997.
M Schwarick (2006):A Tool to analyse Petri net models (in German);BTU Cottbus, Dep. of CS, Master Thesis, September 2006.
A Fischer (2009)Reachability graph analysis of time-dependent Petri nets (in German);BTU Cottbus, Dep. of CS, Master Thesis, Oktober 2009.
A Franzke (2009)Charlie 2.0 - a multi-threaded Petri net analyzer, BTU Cottbus, Dep. of CS, Master Thesis, December 2009
J Wegener, M Schwarick, M Heiner (2011)A Plugin System for Charlie;Proc. CSP 2011, Biaystok University of Technology, 531-554, September 2011.
Marcie
A Noack (1999)A ZBBD Package for Efficient Model Checking of Petri Nets (in German); BTU Cottbus, Dep. of CS, Major Individual Project, 1999.
A Tovchigrechko (2008)Efficient symbolic analysis of bounded Petri nets using Interval Decision Diagrams;PhD thesis, BTU Cottbus, Dep. of CS, October 2008.
M Heiner, M Schwarick, A Tovchigrechko (2009)DSSZ-MC – A Tool for Symbolic Analysis of Extended Petri Nets;Proc. PETRI NETS 2009, Paris, Springer, LNCS, volume 5606, pages 323–332, June 2009
M Schwarick, A Tovchigrechko (2010)IDD-based model validation of biochemical networks;Theoretical Computer Science, July 2010.
M Schwarick, C Rohr, M Heiner (2011)MARCIE - Model checking And Reachability analysis done effiCIEntly;Proc. QEST 2011, Aachen, Germany, IEEE CS Press, pages 91–100, September 2011.
M Heiner, C Rohr, M Schwarick (2013)MARCIE - Model checking And Reachability analysis done effiCIEntly;Proc. PETRI NETS 2013, Milano, Springer, LNCS, volume 7927, pages 389–399, June 2013.
M Schwarick (2014)Symbolic on-the-fly analysis of stochastic Petri nets;PhD thesis, BTU Cottbus, Dep. of CS, June 2014.