obeo thales@md day2011

Post on 19-Jan-2015

1.642 Views

Category:

Technology

0 Downloads

Preview:

Click to see full reader

DESCRIPTION

 

TRANSCRIPT

1

Model-driven Architecture building

for constrained Systems

Jean-Luc Voirin

THALES

From Requirements allocation to Architecture

Mastering

late discovery of design issues during IVVQ

Requirements

« Requirement to Boxallocation »

Non-functionalConstraints

Architecture building & Justification

OperationalNeed

FunctionalNeed

Reqs

Trade-off

Architecture

Early validation of the Architecture Mastering and optimisation of

product & IVVQ

3

ARCADIA Method

Non-functionalConstraints

Non-functionalViewpointsTrade-off

Exte

nsi

ons

Transition to sub-system,SW,HW

Adds brand new multi-viewpoint non-functional analysis

ARCADIA

OperationalAnalysis

FunctionalAnalysis

LogicalArchitecture

PhysicalArchitecture

EPBS

Integrates & enriches them to a much wider scope

A model-based Architecture engineering approach

DoDAF, NAF

SysML,

UML2

AADL

thus allowing architecturecollaborative validation& interoperability through standards

Starts from standards

IEEE STD 1471

Method Steps: formalising & sharing Need & Solution

Requirements

System

Functional & NF NeedModel

Logical ArchitectureModel

Physical ArchitectureModel

Operational

AnalysisModel

Product Breakdown

User Need

System Need

Notional Solution

Final Solution

Sub-contractors input

What the users of the system need

to accomplish

What the system has to accomplish

for the Users

How the system will work so as to

fulfil expectations

How the system will be developed

& built

Engineering

documents:

SSS, IRS,

OCD, …

Engineering

documents:

SSDD, ICD,

PIDS, SRS,

5

Summary of ARCADIA principles & capability

Safety

Cost & Schedule

Industrial Constraints

Maintainability

Product Policy

Etc…

Interfaces

Functional Analysis

Performance

Automatic Analysis

OperationalUser need

Industrial Stakes

CandidateArchitectures

Validated solution architecture

-Justified- Shared-Capitalised

Sample Example (bird eye): signal, data & info

processing

OperationalAnalysis

Data & Interface Analysis

Logical & Physical Architectures

Functional Analysis

Modelling allows checking solution against Need

Including non-functional constraints

Early Validation: How to check Solution Architecture

Perform a multi-viewpoint trade-off Analysis

Solution

ArchitectureComponents

Evaluation Rules

SharedShared

Architecture Models

Product Line

Managers

Architects

Requirement

Managers

Others…

Safety, Performance,

interface… Engineers

Logistics

Engineers

SW/HW

Designers

& Developpers

Customers

IVVQ

Manager

Architect

Specialty

engineering:safety, perf,

interface, …

Product Line

Manager etc…

Sub-

contractors

ViewPoints

FunctionsSafety

Security

Performance

Interfaces

IVVQ, Product Line, Cost…

weight

Cost…

Safety

Security

Performance

Interfaces

Product line

Integration/V/V

8

Performance non-functional ViewpointNon-

functionalConstraints

Non-functionalViewpointsTrade-off

Exte

nsi

ons

Transition to sub-,SW,HW

FunctionalAnalysis

LogicalArchitecture

PhysicalArchitecture

EPBS

OperationalAnalysis

1° Automatic analysis:-CPU overloaded-Bus overloaded-Latency in excess

2° Causal analysis-Tool locates problems-Quantitative analysis

3° Architecture

improvement:-Lighter protocols-Higher bandwidth-Hardware processing

9

3° Architecture

improvement:-Second redundant server

2° Causal analysis-Tool locates problems-Failure propagation algorithm

Safety non-functional ViewpointNon-

functionalConstraints

Non-functionalViewpointsTrade-off

Exte

nsi

ons

Transition to sub-,SW,HW

FunctionalAnalysis

LogicalArchitecture

PhysicalArchitecture

EPBS

OperationalAnalysis

1° Automatic analysis:-Rule: “No single source for major failure condition”-Not met for video

2° Causal analysis-Tool locates problems

10

Multi-Viewpoints Trade-offNon-

functionalConstraints

Non-functionalViewpointsTrade-off

Exte

nsi

ons

Transition to sub-,SW,HW

FunctionalAnalysis

LogicalArchitecture

PhysicalArchitecture

EPBS

OperationalAnalysis

Confrontation rules for multi-viewpoints trade-off

Field of application

Initially developped to target

signal processing, data processing,

information systems

Appears to be applicable to much larger field

e.g. electrical power systems

Thermal systems

MDDay 2011 11

12

Command & Control Model (coming soon)

Interfacing &computing units…

Data

Example: Three interleaved, multi-physics Models

Thermal ModelTurbines, Compressors, heat exchangers, valves…

Thermal Power

-Conditioning

-Pressurisation

-Equipments Cooling

Power Model

Electrical Generators, Loads, converters, bus bars…

Electrical power-Generation

-Distribution

13

Power / thermal Model dedicated Viewpoints

Viewpoints

Solution

Architecture

Viewpoints

Solution

Architecture

Power & Thermal performancedepending on flight phase consumption, incl. Overloaded components detection based on power computation, linked to thermal Model

Safety / Integrityincl. Failure containment, redundancy rules & analysis, failure scenarios & propagation, monitoring efficiency, shielding…

Reliability & Availabilityincl. Reliability computing, reconfiguration issues, flight delay…

And also: Mass, Cost, Reliability…

Spatial (3D) arrangementEarly identification of spatial arrangement constraints impacting the architecture

14

Conclusion: Return of Experiment

Early validation of key architectural aspectsLess rework in design & production

Efficient decision making regarding complex but necessary architectural trade-off

Improved quality of architecture

Support to negotiation and compromise between stakeholders

Ability to capitalise both product definition and know-how

15

Conclusion: Return of Experiment

Support to interoperation with Customers & Suppliers

Interoperability with and beyond Standards Support

Benefits and guidelines for foreign tools integration,esp. simulation and specialty tools

Ability to adjust modelling effort:scaled/focused on major engineering issues for return on invest

Without exhaustive modelling

ARCADIA Support by Melody Advance Modeller

Melody Advance solves Weaknesses of COTS* or OSS*:

ARCADIA Method support & guidance for modelling

• Method Steps, encyclopedia, rules, diagrams…

User-oriented Semantics

• Engineer concepts rather than abstract/profiled language

Support for « modelling in the large »

• Performance on large models, ergonomy, modelling aids…

Support for Viewpoint extensions, modelling & analysis

• Model extensions, diagrams extensions, viewpointmaangement…

« semantic » Import/export capabilities (excel, SysML, AF, …)Yet ARCADIA is also deployed using other tools

• Excel/Access, Rhapsody, System Architect/DoDAF…

• with reduced capabilities, however

* COTS: Commercial Off The Shelf

* OSS: Open Source Software

top related