Project P erts2012
-
Upload
adacore -
Category
Technology
-
view
6.165 -
download
3
description
Transcript of Project P erts2012
![Page 1: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/1.jpg)
Compiling Heterogeneous Models: Motivations and Challenges
Matteo Bordin, Tonu Naks, Marc Pantel and Andres Toom
ERTS2 2012 – February 1st-3rd, Toulouse
![Page 2: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/2.jpg)
The topic of this presentation
2
![Page 3: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/3.jpg)
Original scope of the project (2010)
• Evolve/Productize GeneAuto technology
– A model compiler for Simulink, Stateflow and Scicos
– ITEA project, 2005-2008 (see ERTS 2008/2010)
– Mostly developed by IB Krates, IRIT and Alyotech
– Validated in 2009-2010 by Airbus and Astrium (see ERTS 2010)
• Emphasis on qualification in a DO-178 context
• Small, focused project
3
![Page 4: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/4.jpg)
4
Fast forward to October 2011…
![Page 5: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/5.jpg)
5
avioncis
space automotive
Industrial users
Academia Tech providers
![Page 6: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/6.jpg)
Model-based integration Qualifiable code generators
Tunable & Open-Source
![Page 7: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/7.jpg)
Positioning the P toolset
7
Specification Model
Design Model
Src Code
Executable
• Compliance • Robustness
• Compliance • Traceability • Accuracy • Consistency
•Conformity •Verifiability •Accuracy & Consistency •Compatibility
![Page 8: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/8.jpg)
8
Or, in simpler terms…
… the model compiler for
![Page 9: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/9.jpg)
9
Technical approach
![Page 10: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/10.jpg)
10
Architecture
Behaviour
Control algorithms Signal processing
System and platform
System Engineer Software Engineer Automation Engineer
importer importer importer
Integration verification
Simulink® Stateflow®
Ada C VHDL
Code Generation
Formalism
Links between models
subset subset subset
![Page 11: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/11.jpg)
11
Project P genealogy
![Page 12: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/12.jpg)
12
SPaCIFY
![Page 13: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/13.jpg)
Ultimate goals: a tool provider perspective
• To put an end to segregation
– Between system, software, control engineers
• To deploy a lean/agile qualification process
– Qualify as you develop, easy re-qualification
• To build an economically viable ecosystem
– For both tech providers and service companies
13
![Page 14: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/14.jpg)
Ultimate goals: a scientist perspective
• To put an end to segregation between
– System, software and control engineers
– Industrial and academic partners
– Formal methods and engineering tools
• To define a common intermediate language
– Providing a well defined semantics
– To connect analysis, synthesis and verification tools based on
different elementary formal languages
– To compare, improve and combine these tools
– To connect with industrial modeling tools
14
![Page 15: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/15.jpg)
15
The challenges
![Page 16: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/16.jpg)
Semantic convergence of DSLs (I)
• P formalism is not the union of all input languages
• Example: functional behavior
– Synchronous state machines: Simulink, Stateflow
– Asynchronous state machines: UML, AADL-BA
– Execution protocols: AADL, UML
• Example: architectural modeling
– SysML/UML/MARTE
– AADL/ARINC-653
– Autosar/East-ADL2 (TIMMO)
16
![Page 17: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/17.jpg)
Example: semantic convergence of behavioral DSLs
17
Formalism
![Page 18: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/18.jpg)
Semantic convergence: Current proposal
• Derived mainly from TOPCASED-AADL and GeneAuto
• Strong separation between
– Function (derived from Synchronous MoC)
Black box: no interaction with architecture during execution
Synchronous w.r.t. Architecture
Can be internally concurrent
– Architecture (derived from AADL MoC)
Clock management
Synchronous/Asynchronous communication
Distribution/Process/Thread management
18
![Page 19: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/19.jpg)
Verification of model-based integration
• Well beyond typing
– Clock constraints
– Synchronization
– Hard real time properties
• Allocation of behavior
– Extra functional properties are assumed to hold in simulation
– Are they respected by the architecture?
• Support for the synthesis of software architecture
– Logical and physical
19
![Page 20: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/20.jpg)
Flexible & Open Qualification • Support for future input DSLs
– Separable qualification evidence for:
Importers
Code generators
• Support for new code generation strategies
– Separable qualification evidence for:
Target independent model transformations
“Printing phase”
• Toolkit shipping with qualification infrastructure
– To allow incremental re-qualification
20
![Page 21: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/21.jpg)
21
Current state at T0+4
![Page 22: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/22.jpg)
22
Figuring out the consortium agreement
![Page 23: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/23.jpg)
Synchronizing with on-going projects • CHESS, VERDE
– Inspiration for component model and architecture modeling
• MBAT
– Model verification of the P formalism
• OPENCOSS & Safecer
– Tooling for qualification artifact management
• TASTE
– AADL tooling
• openETCS
• CESAR
• OPEES/PolarSys
• . . . 23
![Page 24: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/24.jpg)
Technical work
• Migrating GeneAuto technology
– Due to unclear DO-178 qualifiability of technology
• DO-178C –compliant qualification data
– Migrating GeneAuto data
– User-TOR, Developer-TOR, …
• Experimenting the Open-DO collaborative platform
– Interactive wiki instead of emails
– Wiki pages instead of Word for deliverables
24
![Page 25: Project P erts2012](https://reader033.fdocuments.in/reader033/viewer/2022051818/54bd0ad44a795959428b459f/html5/thumbnails/25.jpg)
25