CESAR - Cost-efficient methods and processes for safety relevant embedded systems CESAR – APPROACH & MEASUREMENT OF OBJECTIVES Bourrouilh Quentin, Kundner Ingrid / AVL
CESAR - Cost-efficient methods and processes for safety relevant embedded systems
CESAR – APPROACH &
MEASUREMENT OF OBJECTIVES
Bourrouilh Quentin, Kundner Ingrid / AVL
Bourrouilh, Kundner / AVL
10-06-302
MOTIVATION – AUTOMOTIVE
1st Vehicles a decade ago� A few embedded systems per
vehicle
Vehicles nowadays� Up to a few hundreds of
computing devices per vehicle
� Multiple networks per vehicle
Advantage: � Safety-critical embedded
systems have been key innovation drivers
� E.g. by-wire systems
Disadvantage: � Enormous complexity is
challenging industry (automotive, aerospace, rail, automation)
� Increasing costs
� Affected product quality �safety-critical
PAST TODAY
FUTURE ?
Bourrouilh, Kundner / AVL
10-06-303
MOTIVATION - AEROSPACE
Engine equipment and parts
• Integrated engine control systems
• Power transmissions
• Engine modules and components
• Composite engine parts
Engines
• CFM56 family (50/50 with GE)
• SAM146 engine for the Russian
Regional Jet (50/50 with NPO Saturn)
• Participation in programs: CF6, GE90,
GE90-115B, GP7000, PW4000, AS900, CF34
Landing & braking systems
• Landing gear for all types of aircraft
• Braking/landing control systems
• Wheels and carbon brakes
• Control systems and hydraulics
• Maintenance, repair and overhaul
Aircraft equipment
• Network server systems
• Back-up flight control
• Secure data link
• Cockpit control systems
• Electrical wiring systems
• Aircraft condition monitoring systems
Engine services
• Maintenance, repair and overhaul
• Engine testing and test equipment
• Composite aerostructures
• Auxiliary power units
• Hydraulic systems
• Sensors and actuators
• Ventilation/filtration
• Inertial references
Nacelles
Nacelles and components
(thrust reversers,…)
Bourrouilh, Kundner / AVL
10-06-304
CESAR OBJECTIVES
Improvement of processes and methods for safety-critical embedded systems development
Development of the reference technology platform (RTP) for safety-critical embedded systems development
Motivate technology providers and SMEs to contribute to the RTP
Bring innovations in two most improvable engineering disciplines� Requirements engineering
� Component-based engineering
Bourrouilh, Kundner / AVL
10-06-305
CESAR OBJECTIVES
Reduce costs for the development of safety critical systems while ensuring the quality and safety properties
Covering the entire System Engineering (SE) disciplines:
� Improve Requirements engineering (RE)
� Improve Component-based development (CBD) and extend it with multi views and multi criteria
� Combine the improved RE (e.g. RMM) and Design System Engineering (incl. CBD)
� A close collaboration between RE and CBD is necessary to satisfy the ambitious CESAR goals
� Only an integration of these disciplines accompanied with an adequate tool support into a seamless tool chain (CESAR RTP) can unleash the full potential of the CESAR approach
Bourrouilh, Kundner / AVL
10-06-306
CESAR APPROACH
� A main objective is to reduce costs
� CESAR will answer this by improving the SE (System Engineering) discipline
� SE considers the necessary combination of RE, CBD, Safety and PLE
Innovations in tools and methods
RE (Requirements Engineering)
SCADE
Simulink
Tool A
Rhapsody
Tool B
RT-Builder
Tool X
Enovia
Reqtify
Tool Y
Analysis X
Analysis Y
Innovations in tools and methods
CBD (Component Based Development)
Safety
PLE
Bourrouilh, Kundner / AVL
10-06-307
CESAR PROJECT STRATEGY
� Industrial driven
� Multi-domain approach
� 3 Innovation Cycles
� Industrial needs (requirements)
� Solutions (don’t reinvent the
wheel) provided by technical
Subprojects + common bottom-
up and top-down approach
� Results (Technical Items) back
to Domain SPs (integrated or
not integrated in the RTP)
� Evaluation and feedback
through the end users Pilot
Applications
Bourrouilh, Kundner / AVL
10-06-308
CESAR FACTS & FIGURES
� Consortium
� 55 Partners
� Further Assisting Parties
� Project Performing
� Duration: 3 years
� Start: 01.03.2009
� Manpower
� Effort: 5124 MM ~ 142 MY/Y
� Project Figures
� Total Budget: 58.535.000 €
� Total Funding: 28.317.000 €
Bourrouilh, Kundner / AVL
10-06-309
CESAR PROJECT STRUCTURE
AVL
OFFIS
SIEMENS
A-F
VO
LV
O
MB
AB
BEADS-IW
SAGEM
PSB
TPC
AVL
SAGEM
Technical SPs are lead by an academic and an indusrial partner
Bourrouilh, Kundner / AVL
10-06-3010
SP1: REFERENCE TECHNOLOGY PLATFORM
� Realization of RTP for safety-critical embedded systems development
Task Force Safety and Diagnosability� Appropriate and efficient
approaches to developmentand validation
Task Force Product Lines� Insurance of a consistent
product line engineering process throughout development cycle
Bourrouilh, Kundner / AVL
10-06-3011
SP2 & SP3
SP2: Requirements engineering� Improve RE methods and
processes
SP3: Component-based development� Facilitation of system
exploration
� Provide means for incremental verification, validation and certification
Bourrouilh, Kundner / AVL
10-06-3012
SP2: REQUIREMENTS ENGINEERING
WHY: � Proper requirements engineering essential in safety projects
� Provide evidence for achievement of functional safety
� Provide extensive documentation for certification
HOW:� Improve current process dealing with requirements engineering and management
� Develop a formalized multi criteria requirement specification language called RSLand a corresponding RMM:
� High level of abstraction: requirements in textual form
� Low level of abstraction: specified in formal notation
� Tool-independent exchange format
� Ensure complete traceability, completeness and consistency from concepts to products
� Support validation of formal multi-criteria requirements with respect to consistency and completeness
� Implementation of improved requirements engineering methods and tools as well providing corresponding guidelines
Bourrouilh, Kundner / AVL
10-06-3013
SP3: COMPONENT BASED DEVELOPMENT
WHY:� Safety-critical embedded systems development is costly and
reuse is difficult
HOW:� Focus of SP3 is on system architecture and system detailed
design which relies on component based design
� Support integration of multiple views and multi-criteria in design space exploration
� E.g. functionality, hardware, safety, performance
� Main activities
� Modeling languages and validation techniques
� Methodologies for specifying and validating system architectures as well detailed system design made from components
� Tool chains specification and tool implementation
Bourrouilh, Kundner / AVL
10-06-3014
SP5, SP6, SP7: DOMAIN SPs
Provision of industrial needs� Pilot Applications
� State of practice
� Safety Design Process
� Requirements Description
Support of existing and evolving domain standards:� ISO 26262 Functional Safety
� IEC61508
� DO178
� ENxxx
Inputs for technical subprojects and support for evaluation and validation
PA examples:� Power train control unit for hybrid
vehicle
� Engine control system for aircraft
Bourrouilh, Kundner / AVL
10-06-3015
CESAR MULTI DOMAIN APPROACH
4 DOMAINS - INDUSTRIAL PARTNER
� Automotive (SP5)
� Aerospace (SP6)
� Automation & Rail (SP7)
Bourrouilh, Kundner / AVL
10-06-3016
ACHIEVEMENTS YEAR 1
2nd JU Review on the 17th and 18th of May
Very successful and green light to continue
Examples of Top-Level achievements so far:� Cartography of Pilot Applications (PA) and Technical Innovations
(TI)
� SP1: RTP V1.0 is available and first evaluation starts
� SP2: « RE process » techniques (RSL, RMM) and methods have been identified
� SP3: Selection and identification of methodologies and data model to support the multiviews modelling
Bourrouilh, Kundner / AVL
10-06-3017
CESAR MEASUREMENT OF OBJECTIVES
Bourrouilh, Kundner / AVL
10-06-3018
CESAR OBJECTIVES
Scope
Short name
Area Quantitative Objective
Aero
sp
ace
Au
tom
otiv
e
Rail
Au
tom
atio
n
Innovation cutting cost by 30-50%
Process
Introduce in each domain at least one significant innovation in design, integration or validation process, clearly supported by CESAR, acceptable across the supply chain and to certification authorities when appropriate,
resulting in overall reduction of development time or effort, between 30% and 50%, depending on the domain
x x x x
Reduce effort of revalidation by 50%
ProcessDemonstrate, at least in one domain, a reduction by 50% of the effort of re-validation and re-certification after change, the process being acceptable
across the supply chain if appropriate, and to certification authorities
x †[1] x x
100%
complexity increase, 20% effort reduction
ProductProcess
Demonstrate, at least in one domain a 100% complexity increase of the product with 20% engineering effort reduction
x
Product capability improvement without increased cost
ProductIntroduce in each domain at least one major product capability improvement
clearly related with CESAR, without impact on recurring costx x x x
[1] Not applicable for this domain (no certification authority)
Bourrouilh, Kundner / AVL
10-06-3019
FROM OBJECTIVES TO TECHNICAL ITEMS
CESAR has to track the overall objectives -systematic breakdown from objectives to solutions
Bourrouilh, Kundner / AVL
10-06-3020
3 STEP PROCESS
� 1. Creation of Baseline
� Metrics definition
� Creation of a harmonized development process, valid for all CESAR domains (5 Phases) with domain-specific activities located in each phase
� 2. Mapping of Pilot Applications to Baseline
� CESAR Cartography
� 3. Measure Impact of Technical Innovations
� Evaluation on 4 different perspectives including their own metrics:
� Efficiency (%)
� Quality (%)
� Complexity (%)
� Cost / Effort (%)
Bourrouilh, Kundner / AVL
10-06-3021
CREATION OF BASLINE: METRICS DEFINITION (1/5)
Scope
Short name
Area Quantitative Objective
Aero
sp
ace
Au
tom
otiv
e
Rail
Au
tom
atio
n
Innovation cutting cost by 30-50%
Process
Introduce in each domain at least one significant innovation in design, integration or validation process, clearly supported by CESAR, acceptable across the supply chain and to certification authorities when appropriate,
resulting in overall reduction of development time or effort, between 30% and 50%, depending on the domain
x x x x
Reduce effort of revalidation by 50%
ProcessDemonstrate, at least in one domain, a reduction by 50% of the effort of re-validation and re-certification after change, the process being acceptable
across the supply chain if appropriate, and to certification authorities
x †[1] x x
100%
complexity increase, 20% effort reduction
ProductProcess
Demonstrate, at least in one domain a 100% complexity increase of the product with 20% engineering effort reduction
x
Product capability improvement without increased cost
ProductIntroduce in each domain at least one major product capability improvement
clearly related with CESAR, without impact on recurring costx x x x
[1] Not applicable for this domain (no certification authority)
CostComplexity
Cost
Complexity
Bourrouilh, Kundner / AVL
10-06-3022
CREATION OF BASLINE: METRICS DEFINITION (2/5)
Cost
QualityComplexityEfficiency
×=
Efficiency
QualityComplexityCost
×=
In other words...
Bourrouilh, Kundner / AVL
10-06-3023
CREATION OF BASELINE: METRICS DEFINITION (3/5)
� A set of metrics (measurable) and indicators (qualitative) have been issued. This list is not exhaustive and will be refined after 1st RTP evaluation tests depending on stated measurability.
WG-Objectives
Bourrouilh, Kundner / AVL
10-06-3024
CREATION OF BASELINE: METRICS DEFINITION (4/5)
# Area Objective Associated goal(s)
A1 Process
Introduce in each domain at least one significant innovation in design, integration or validation process, clearly supported by CESAR, acceptable across the supply chain and to certification authorities when appropriate, resulting in overall reduction of development time or effort, between 30% and 50%, depending on the domain
Cost reduction Efficiency
A2 Process
Demonstrate, at least in one domain, a reduction by 50% of the effort of re-validation and re-certification after change, the process being acceptable across the supply chain if appropriate, and to certification authorities
Efficiency Quality
A4 Product
Introduce in each domain at least one major product capability improvement clearly related with CESAR, without impact on recurring cost
Quality
B4 RTP
Reduce by 50% the cost of integration, configuration, deployment, and maintenance of appropriate tool-chains for all major actors in the supply chain involved in the project
Cost reduction
B5 RTP
Improve the quality of support tools by targeting a level of maturity at EIS consistent with the needs of critical systems engineering: reduce by 50% the number of Problem Reports on support tools after their delivery to system engineering teams.
Quality Efficiency
Coverage versus CESAR Objectives (SP6 related)
Bourrouilh, Kundner / AVL
10-06-3025
CREATION OF BASELINE: METRICS DEFINITION (5/5)
� Fact: All goals are linked together ! The goals are only different views to the entire development to derive metrics:
� Goal 1 Efficiency :
� Sub-goal 1.1: Reduce development cycle time
� Sub-goal 1.2: Find defects as early as possible
� Sub-goal 1.3: Improve requirements stability
� Goal 2 Quality :
� Sub-goal 2.1: Improve reliability
� Sub-goal 2.2: Improve system model maintainability
� Sub-goal 2.3: Improve system’s documents quality
� Goal 3 Complexity :
� Sub-goal 3.1: Improve management of configurability
� Sub-goal 3.2: Improve verification ability
� Sub-goal 3.3: Improve handling of change
� Sub-goal 3.4: Improve system modelling capabilities
Bourrouilh, Kundner / AVL
10-06-3026
GOAL 1: IMPROVE EFFICIENCY 1/3
� Sub-goal 1: Reduce development cycle time
� Measurement goal: Analyse the impact of RTP tool usage on the development time.
� Questions:
� Does RTP tool reduce development time?
� Indicators:
� Distribution of the normalised elapsed time (normalised by total size) of each development phase and for the whole development process.
� Measures: For each Pilot applications:
� Elapsed time for each development phase
� Total development effort
Bourrouilh, Kundner / AVL
10-06-3027
3 STEP PROCESS
� 1. Creation of Baseline
� Metrics definition
� Creation of a harmonized development process, valid for all CESAR domains (5 Phases) with domain-specific activities located in each phase
� 2. Mapping of Pilot Applications to Baseline
� CESAR Cartography
� 3. Measure Impact of Technical Innovations
� Evaluation on 4 different perspectives including their own metrics:
� Efficiency (%)
� Quality (%)
� Complexity (%)
� Cost / Effort (%)
Bourrouilh, Kundner / AVL
10-06-3028
CREATION OF BASELINE: HARMONIZATION
� Harmonization of Development process for all CESAR Domains
� 5 Phases:
� User’s needs
� System Design & HW/SW Specification
� HW/SW Design Coding & Build Unit Testing
� System Integration Testing
� Certification / Qualification activities
� Support Activities
� Domain specific activities
Bourrouilh, Kundner / AVL
10-06-3029
CREATION OF BASELINE: EFFORT ALLOCATION (1/2)
� Allocation of development effort to harmonized development process
� Development effort before impact of CESAR Technical Innovations
Bourrouilh, Kundner / AVL
10-06-3030
CREATION OF BASELINE: EFFORT ALLOCATION (2/2)
Requirement capture and elicitation
Requirement translation in formal
language
Safety assessmentRequirement management
Validation
Document generation
HW SW integrationRequirement managementVerification
Document generation
Requirement managementDocument generation
Requirement managementSW development and reuseHW development and reuse
Safety assessmentVerificationDocument generation
Requirement management
System architecture definition
Safety assessment Prototyping
Automatic test generation
Validation
HW/SW Specification
Document generation
Support Activities : Collaborative work
Configuration management
Phase 1
User's needs
Phase 2
System Design
& HW/SW
Specification
Phase 3
HW/SW Design,
Coding & Build,
Unit Testing
Phase 4
System Integration
Testing
Phase 5
Certification /
Qualification
activities
X %X %
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
Phase 1 Phase 2 Phase 3 Phase 4 Phase 5
X %X %
X %X %
X %X %
X %X %
Overall baseline represents 100% of cost/effort repartition on our development process
Main activity level allows :• Finest baseline• Mapping of main activities addressed for each PA �baseline for each pilot application.
Bourrouilh, Kundner / AVL
10-06-3031
3 STEP PROCESS
� 1. Creation of Baseline
� Metrics definition
� Creation of a harmonized development process, valid for all CESAR domains (5 Phases) with domain-specific activities located in each phase
� 2. Mapping of Pilot Applications to Baseline
� CESAR Cartography
� 3. Measure Impact of Technical Innovations
� Evaluation on 4 different perspectives including their own metrics:
� Efficiency (%)
� Quality (%)
� Complexity (%)
� Cost / Effort (%)
Bourrouilh, Kundner / AVL
10-06-3032
MAPPING OF PILOT APPLICATION TO BASELINE
PAs
Phases & Activities
Mapping
� Not all PA cover the entire development process but: the sum of all PA covers the whole System Engineering process.
Bourrouilh, Kundner / AVL
10-06-3033
3 STEP PROCESS
� 1. Creation of Baseline
� Metrics definition
� Creation of a harmonized development process, valid for all CESAR domains (5 Phases) with domain-specific activities located in each phase
� 2. Mapping of Pilot Applications to Baseline
� CESAR Cartography
� 3. Measure Impact of Technical Innovations
� Evaluation on 4 different perspectives including their own metrics:
� Efficiency (%)
� Quality (%)
� Complexity (%)
� Cost / Effort (%)
Bourrouilh, Kundner / AVL
10-06-3034
MEASUREMENT OF TI-IMPACT
� Parallel Performance or
� Measurement versus Baseline
Bourrouilh, Kundner / AVL
10-06-3035
BACKUP
Bourrouilh, Kundner / AVL
10-06-3036
GOAL 1: IMPROVE EFFICIENCY 1/3
� Sub-goal 1: Reduce development cycle time
� Measurement goal: Analyse the impact of RTP tool usage on the development time.
� Questions:
� Does RTP tool reduce development time?
� Indicators:
� Distribution of the normalised elapsed time (normalised by total size) of each development phase and for the whole development process.
� Measures: For each Pilot applications:
� Elapsed time for each development phase
� Total development effort
Bourrouilh, Kundner / AVL
10-06-3037
GOAL 1: IMPROVE EFFICIENCY 2/3
� Sub-goal 2: Find defects as early as possible
� Measurement goal: Analyse the system defects introduction and detection in order to identify opportunities for reducing cost.
� Question:
� How much effort is spent due to discovering defects late in time?
� Indicators:
� Effort spent in fixing failures by introduction phase
� Effort spent in fixing failures by discovering phase.
� Relative effort spent in corrective maintenance (relative to development effort) vs. development effort
� Measures: For each pilot application:
� No. of failures
� For each failure in each pilot application:
� Total effort for fixing a failure
� Phase where failure introduced
� Phase where failure found
Bourrouilh, Kundner / AVL
10-06-3038
GOAL 1: IMPROVE EFFICIENCY 3/3
� Sub-goal 1: Improve requirements stability
� Measurement goal: Assess requirements requests for change in order tounderstand their impact on schedule completion date.
� Questions:
� Does the number of changes to requirements decrease with time?
� How many requirements change requests were received in each development phase and for the whole pilot application?
� What are the main reasons for requirements change requests?
� Indicators:
� Profile of current and cumulative No. of requirements changes.
� Distributions of requirements change requests during each development phase.
� Distributions of requirements change requests according to reason.
� Measures: For each pilot application
� No. of requirements change requests.
� No. of requirements change requests per development phase
� No. of requirements change requests per reason of change