Model-Based System Architecture
Description
AN UP-TO-DATE EXPLORATION OF THE NEWEST STANDARDS AND BEST PRACTICES IN SYSTEM ARCHITECTING
In the newly revised Second Edition of Model-Based System Architecture, a team of expert engineers deliver a detailed and authoritative review of the practice of system architecture in organizations that use models to support the systems engineering process. In the book, readers will find introductions to the fundamentals of architecting systems and using models to assist the architecting process.
The latest edition offers refreshed content based on ISO 15288:2015 and a renewed focus on the role of the system architect. New chapters on systems-of-systems, and cyber-physical systems, and system architect tools offer guidance to practicing professionals on how to apply the presented concepts in the real-world.
In addition to the latest definitions of the architecture governance and evaluation processes described in ISO 42020 and 42030, the book provides:
* A thorough introduction to the value of systems architecting, definitions of system architecture, and model-based system architecture
* Comprehensive explorations of model governance, architecture descriptions, patterns, and principles, and the roles of typical architecture stakeholders
* Practical discussions of Agile approaches to systems architecture, the FAS Method, and architecture frameworks
* In-depth examinations of systems architecting work and necessary soft skills for systems architects
* Modeling of system architectures with SysML including a brief overview of SysML v1 and an outlook to SysML v2
Perfect for system architects and system engineers, Model-Based System Architecture will also earn a place in the libraries of students and researchers studying functional architectures.
More details
Other editions
Additional editions
Persons
TIM WEILKIENS is Executive Board Member of Oose, a German engineering consultancy, and a co-author of the SysML specification.
JESKO G. LAMM is a Senior Systems Engineer in the European hearing healthcare industry.
STEPHAN ROTH is a coach, consultant, and trainer for systems and software engineering at oose. He is a certified systems engineer (GfSE)®- Level C.
MARKUS WALKER is Elevator System Architect in the CTO Division at Schindler Elevator. He is an INCOSE Certified Systems Engineering Professional (CSEP).
Content
Foreword xv
Preface xvii
About the Companion Website xxi
1 Introduction 1
2 An Example: The Scalable Observation and Rescue System 5
3 Better Products - The Value of Systems Architecting 9
3.1 The Share of Systems Architecting in Making Better Products 9
3.2 Benefits that can be Achieved 10
3.2.1 Benefit for the Customer 10
3.2.2 Benefit for the Organization 12
3.3 Benefits that can be Communicated Inside the Organization 14
3.4 Beneficial Elements of Systems Architecting 15
3.5 Benefits of Model-Based Systems Architecting 16
4 Systems, Systems of Systems, and Cyber-Physical Systems 17
4.1 Definition of "System" 17
4.1.1 System Elements 19
4.1.2 System Context 20
4.1.3 System Characteristics 21
4.1.4 Purpose 22
4.1.5 System Evolution 23
4.2 Definition of "System of Systems" 23
4.3 Definition of "Cyber-Physical System" 26
4.4 Composition of a "Cyber-Physical System of Systems" 27
5 Definition of System Architecture 31
5.1 What Is Architecture? - Discussion of Some Existing Definitions 31
5.2 Relations Between Concepts of "System," "Architecture," and "Architecture Description" 33
5.3 Definition of "Architecture" 35
5.3.1 Interactions 36
5.3.2 Principles 37
5.3.3 Architecture Decisions 37
5.4 Functional and Physical Architecture 37
5.5 Taxonomy of Physical Architectures 39
5.5.1 Logical Architecture 40
5.5.2 Product Architecture 41
5.5.3 Base Architecture 41
5.6 Architecture Landscape for Systems 41
5.6.1 System Architecture 42
5.6.2 System Design 43
5.6.3 Discipline-Specific Architecture and Design 44
6 Model-Based Systems Architecting 45
7 Model Governance 51
7.1 Overview 51
7.2 Model Governance in Practice 52
8 Architecture Description 57
8.1 Architecture Descriptions for Stakeholders 58
8.2 Definition of "Architecture Description" 60
8.2.1 Architecture Viewpoints 62
8.2.2 Architecture Views 65
8.2.3 Architecture Decisions 67
8.2.4 Architecture Rationales 69
8.3 How to Get Architecture Descriptions? 69
8.3.1 Model-Based Vision 69
8.3.2 Forms and Templates 71
9 Architecture Patterns and Principles 75
9.1 The SYSMOD Zigzag Pattern 76
9.2 The Base Architecture 82
9.3 Cohesion and Coupling 85
9.4 Separation of Definition, Usage, and Run-Time 87
9.5 Separate Stable from Unstable Parts 89
9.6 The Ideal System 89
9.7 View and Model 90
9.8 Diagram Layout 92
9.9 System Model Structure 93
9.10 System Architecture Principles 95
9.11 Heuristics 95
9.11.1 Heuristics as a Tool for the System Architect 95
9.11.2 Simplify, Simplify, Simplify: Strength and Pitfall 97
10 Model-Based Requirements Engineering and Use Case Analysis 99
10.1 Requirement and Use Case Definitions 99
10.2 Model-Based Requirements and Use Case Analysis from the MBSA Viewpoint 102
10.2.1 Identify and Define Requirements 103
10.2.2 Specify the System Context 104
10.2.3 Identify Use Cases 105
10.2.4 Describe Use Case Flows 109
10.2.5 Model the Domain Knowledge 110
10.3 The SAMS Method 112
10.3.1 SAMS Method Definitions 113
10.3.2 SAMS Method 114
10.4 Use Cases 2.0 117
11 Perspectives, Viewpoints and Views in System Architecture 119
11.1 Introduction 119
11.2 The Functional Perspective 121
11.2.1 SysML Modeling of Functional Blocks 123
11.2.2 Architecture Views for the System Architect 124
11.2.3 Different Architecture Views for the Stakeholders of Different Functions 124
11.3 The Physical Perspective 125
11.3.1 Logical Architecture Example 126
11.3.2 Product Architecture Example 127
11.4 The Behavioral Perspective 130
11.5 The Layered Perspective 130
11.5.1 The Layered Approach 130
11.5.2 The Layered Perspective in Systems Architecting 132
11.5.3 Relation to the Domain Knowledge Model 134
11.5.4 Architecting the Layers 136
11.5.5 SysML Modeling of Layers 136
11.6 System Deployment Perspective 142
11.7 Other Perspectives 144
11.8 Relation to the System Context 146
11.8.1 Validity of the System Boundary 146
11.8.2 Using the System Context as a Part of the Stakeholder-Specific Views 146
11.8.3 Special System Context View for Verification 147
11.9 Mapping Different System Elements Across Different Levels 148
11.9.1 Functional-to-Physical Perspective Mapping 149
11.9.2 Mapping More Perspectives 153
11.9.3 Mapping Different Levels 153
11.10 Traceability 155
11.11 Perspectives and Architecture Views in Model-based Systems Architecting 155
11.11.1 Creating Different Architecture Views in a Model-Based Approach 155
11.11.2 Using SysML for Working with Different Perspectives and Architecture Views 157
11.11.3 The Importance of Architecture Viewpoints in Model-Based Systems Architecting 159
12 Typical Architecture Stakeholders 161
12.1 Overview 161
12.2 Requirements Engineering 162
12.3 Verification 163
12.4 Configuration Management 166
12.5 Engineering and Information Technology Disciplines 167
12.6 Project and Product Management 171
12.7 Risk Managers 174
12.8 Development Roadmap Planners 174
12.9 Production and Distribution 177
12.10 Suppliers 178
12.11 Marketing and Brand Management 178
12.12 Management 180
13 Roles 185
13.1 Roles 185
13.2 The System Architect Role 186
13.2.1 Objective 186
13.2.2 Responsibilities 186
13.2.3 Tasks 187
13.2.4 Competences 188
13.2.5 Required Skills of a System Architect 188
13.2.6 Required Skills for Model-Based Systems Architecting 190
13.3 System Architecture Teams 190
13.4 System Architecture Stakeholders 192
13.5 Recruiting System Architecture People 192
13.6 Talent Development for System Architects 194
14 Processes 199
14.1 Systems Architecting Processes 199
14.1.1 Overview 199
14.1.2 Example of Generic Process Steps 201
14.1.3 Example of Concrete Process Steps 202
14.1.4 Validation, Review, and Approval in a Model-Based Environment 203
14.2 Design Definition Process 207
14.3 Change and Configuration Management Processes 207
14.4 Other Processes Involving the System Architect 207
15 Tools for the Architect 209
16 Agile Approaches 213
16.1 The History of Iterative-Incremental Approaches 214
16.1.1 Project Mercury (NASA, 1958) 214
16.1.2 The New New Product Development Game (1986) 215
16.1.3 Boehm's Spiral Model (1988) 216
16.1.4 Lean (1945 Onwards) 217
16.1.5 Dynamic Systems Development Method (DSDM, 1994) 219
16.1.6 Scrum (1995) 220
16.2 The Manifesto for Agile Software Development (2001) 221
16.3 Agile Principles in Systems Engineering 223
16.3.1 Facilitate Face-to-Face Communication 223
16.3.2 Create a State of Confidence 224
16.3.3 Build Transdisciplinary and Self-Organized Teams 225
16.3.4 Create a Learning Organization 225
16.3.5 Design, but No Big Design (Up-Front) 226
16.3.6 Reduce Dependencies 227
16.3.7 Foster a Positive Error Culture 228
16.4 Scaling Agile 228
16.5 System Architects in an Agile Environment 230
17 The FAS Method 233
17.1 Motivation 234
17.2 Functional Architectures for Systems 236
17.3 How the FAS Method Works 239
17.4 FAS Heuristics 242
17.5 FAS with SysML 244
17.5.1 Identifying Functional Groups 244
17.5.2 Modeling the Function Structure 246
17.5.3 Modeling the Functional Architecture 249
17.6 SysML Modeling Tool Support 250
17.6.1 Create Initial Functional Groups 251
17.6.2 Changing and Adding Functional Groups 254
17.6.3 Creating Functional Blocks and their Interfaces 254
17.7 Mapping of a Functional Architecture to a Physical Architecture 254
17.8 Experiences with the FAS Method 256
17.9 FAS Workshops 258
17.10 Quality Requirements and the Functional Architecture 259
17.11 Functional Architectures and the Zigzag Pattern 262
17.12 CPS-FAS for Cyber-physical Systems 263
18 Product Lines and Variants 269
18.1 Definitions Variant Modeling 270
18.2 Variant Modeling with SysML 271
18.3 Other Variant Modeling Techniques 276
19 Architecture Frameworks 279
19.1 Enterprise Architectures 280
19.2 Characteristics of System of Systems (SoS) 282
19.2.1 Emergence 283
19.3 An Overview of Architecture Frameworks 285
19.3.1 Zachman FrameworkTM 285
19.3.2 The TOGAF® Standard 286
19.3.3 Federal Enterprise Architecture Framework (FEAF) 288
19.3.4 Department of Defense Architecture Framework (DoDAF) 289
19.3.5 Ministry of Defense Architecture Framework (MODAF) 290
19.3.6 NATO Architecture Framework (NAF) 291
19.3.7 TRAK 292
19.3.8 European Space Agency Architectural Framework (ESA-AF) 293
19.3.9 OMG Unified Architecture Framework® (UAF®) 295
19.4 System Architecture Framework (SAF) 296
Together with Michael Leute 296
19.4.1 SAF and Enterprise Frameworks 296
19.4.2 SAF Ontology 298
19.5 What to Do When We Come in Touch With Architecture Frameworks 298
20 Cross-cutting Concerns 301
20.1 The Game-Winning Nonfunctional Aspects 301
20.2 Human System Interaction and Human Factors Engineering 303
20.3 Risk Management 304
20.4 Trade Studies 305
20.5 Budgets 306
21 Architecture Assessment 307
22 Making It Work in the Organization 313
22.1 Overview 313
22.2 Organizational Structure for Systems Architecting 314
22.3 Recipes from the Authors' Experience 318
22.3.1 Be Humble 319
22.3.2 Appraise the Stakeholders 319
22.3.3 Care About Organizational Interfaces 319
22.3.4 Show that it Was Always There 321
22.3.5 Lead by Good Example 321
22.3.6 Collect Success Stories and Share them When Appropriate 322
22.3.7 Acknowledge that Infections Beat Dictated Rollout 323
22.3.8 Assign the System Architect Role to Yourself 324
22.3.9 Be a Leader 324
23 Soft Skills 327
23.1 It's All About Communication 328
23.1.1 Losses in Communication 329
23.1.2 The Anatomy of a Message 330
23.1.3 Factors Influencing Communication 333
23.1.3.1 The Language 333
23.1.3.2 The Media Used 333
23.1.3.3 Spatial Distance 333
23.1.3.4 Various Connotations of Words 335
23.1.4 The Usage of Communication Aids and Tools 335
23.2 Personality Types 338
23.2.1 Psychological Types by C. G. Jung 338
23.2.2 The 4MAT System by Bernice McCarthy 340
23.3 Team Dynamics 341
23.4 Diversity and Psychological Safety 342
23.4.1 Project Aristotle (Google) 342
23.4.2 Elements of Psychological Safety 343
23.5 Intercultural Collaboration Skills 344
24 Outlook: The World After Artificial Intelligence 347
Appendix A OMG Systems Modeling Language 349
A.1 Architecture of the Language 350
A.2 Diagram and Model 352
A.3 Structure Diagrams 353
A.3.1 Block Definition Diagram 354
A.3.2 Internal Block Diagram 357
A.3.3 Parametric Diagram 361
A.3.4 Package Diagram 362
A.4 Behavior Diagrams 363
A.4.1 Use Case Diagram 364
A.4.2 Activity Diagram 366
A.4.3 State Machine Diagram 369
A.4.4 Sequence Diagram 371
A.5 Requirements Diagram 372
A.6 Extension of SysML with Profiles 374
A.7 Next-Generation Modeling Language SysML v2 376
Appendix B The V-Model 381
B.1 A Brief History of the V-Model or the Systems Engineering Vee 381
B.2 A Handy Illustration but No Comprehensive Process Description 383
B.3 Critical Considerations 385
B.3.1 The V-Model as Process Description 386
B.3.2 The V-Model Does Not Impose a Waterfall Process 386
B.3.3 The V-Model Accommodates Iterations 387
B.3.4 The V-Model Permits Incremental Development 387
B.3.5 The V-Model and Concurrent Engineering 388
B.3.6 The V-Model Accommodates Change 388
B.3.7 The V-Model Permits Early Verification Planning 388
B.3.8 The V-Model Shows Where to Prevent Dissatisfaction 388
B.4 Reading Instruction for a Modern Systems Engineering Vee 389
B.4.1 The Vertical Dimension 389
B.4.2 The Horizontal Dimension 389
B.4.3 The Left Side 389
B.4.4 The Right Side 390
B.4.5 The Levels 390
B.4.6 Life Cycle Processes 390
B.4.7 The Third Dimension 390
Appendix C Glossary 391
C.1 Heritage of the Term "Glossary" 391
C.2 Terms with Specific Meaning 393
References 399
Index 417
3
Better Products - The Value of Systems Architecting
When driving down to the beach in a nice new car, we may enjoy how well this car grabs the road, and if mentally not yet ready for the beach, we might ask ourselves which department in the car company is responsible for the driver's feeling that the car grabs the road. Is it the suspension design unit or the department with all the steering experts? We believe that all these departments alone cannot make us feel that the car grabs the road, because to do so, the car manufacturer has to see the "car as a system, as a collection of things that interact with each other and with the driver and the road" [166]. Systems architecting will ensure that the interactions between components are controlled in a proper way and that components are designed to fit each other.
3.1 The Share of Systems Architecting in Making Better Products
The example of the car that grabs the road was given by J.N. Martin [166]. We extend this example by speculating what would happen if different developers of car components were asked whose merit it is that the car grabs the road. As a reply, maybe the car manufacturer's suspension department and steering experts as well as the tire companies would claim that they are the ones, by making the best possible suspension, steering, tires, etc., but in contrast to this, consider the following example by Russell L. Ackoff: "Suppose we bring one of each of these [many existing types of automobiles] into a large garage and then employ a number of outstanding automotive engineers to determine which one has the best carburetor. When they have done so, we record the result and ask them to do the same for engines. We continue this process until we have covered all the parts required for an automobile. Then, we ask the engineers to remove and assemble these parts. Would we obtain the best possible automobile? Of course not." [5, p. 18].
Since systems architecting is concerned with making components fit together instead of making them "the best" each on its own, we believe that systems architecting is an approach that will help an organization think, develop, produce, and maintain better products.
3.2 Benefits that can be Achieved
When talking of better products, then "better" can have two different meanings:
- More satisfying or even more enjoyable for customers (as shown with the "grabs the road" example).
- More profitable for the organization.
Of course, the first aspect may lead to the second one because products that are perceived well by the customer have the potential to become best-sellers and thus to generate profit for the organization.
It cannot be stated in general whether an organization sees it as most beneficial to minimize development cost, production cost, or maximize user satisfaction or certain quality measures. In the end, trade studies will determine the optimum trade-offs between these different criteria and probably many more. Systems architecting is the activity that will produce well-founded trade studies, because it is right in between the requirements analysis work and the solution space that is framed by the development of different system elements. Figure 3.1 illustrates this based on the example of system-level trade studies across the subsystems of the system-of-interest. Systems architecting can enable a top-down realization of business goals, requirements, quality criteria and product strategies into solutions. This is almost independent from whether a completely new system is developed (a very rare case) or whether an existing system needs to evolve further, by architecting increments to an existing solution. The only difference is that constraints from the existing solution will enter trade studies in systems architecting in the latter case, via the expertise from subsystem development.
It is via this top-down approach that usability, maintainability, reliability, etc. can be designed into the system and that concepts like design for user experience [100] or design to market [261] can be realized.
Figure 3.1 Systems architecting plays a major role in breaking down requirements to solutions and in optimizing solutions, for example, by finding good trade-offs.
3.2.1 Benefit for the Customer
Different kinds of businesses have different customers: While the consumer goods industry targets millions of individual consumers, subcontractors target industries whose suppliers they are. Despite the variety of customers addressed by different industries, we believe that any system developer will increase customer satisfaction with investments into systems architecting.
The good news for industrial customers of subcontractors is: We expect change requests and risk management to work very smoothly on a well-architected system with a proper architecture description in place. For example, we have seen cases in which the impact of a change could be easily analyzed, based on the architecture description, and since the system architecture captures dependencies inside the system, we also expect it to help analyzing how uncertainty in one area of the system may lead to risks in other areas. Being able to manage risk based on the knowledge of the system architecture is indeed a potential sales driver, if we believe in one of the conclusions of J.P. Monat's article "Why Customers Buy" [176]: customers' perception of risk seems to be an important factor in the purchasing decision.
The good news for consumers of mass products is: We expect a properly architected system to have a good chance of working as expected in the market place, because systems architecting can allow for thinking the different modes of operation into it instead of just testing them into it. Furthermore, well-defined interface descriptions may offer a basis for planning the systematic review or testing of component interactions in order to find flaws. This is particularly interesting with regard to discovering rare cases for which interfaces are not prepared. Well-documented architecture also enables continuous reassessment of quality, e.g. by means of reviews. Hence, there is a good basis for preventing flaws introduced by continuous changes during product improvements. This in turn may lower the threshold for continuous changes with the purpose of continuous product improvement.
Last but not least, the properly architected system has a good chance of achieving an attractive cost-benefit ratio, because the organization that developed the system and the production facility producing it could save development and production cost, as it will be explained in Section 3.2.2.
3.2.2 Benefit for the Organization
The standard ISO/IEC/IEEE 42010:2011 [114] declares that each system exhibits an architecture. In other words, every system development produces system architecture. The question is whether the system architecture evolves implicitly or whether it is explicitly defined.
The precondition for the benefits to be discussed is an explicit way of systems architecting, where system development and its strategic planning explicitly involves system architecture processes. In case these are not yet established, there will be an initial investment into architecture work, before the organization can harvest the expected benefits.
Once an organization has established systems architecting as an integral part of system development, it should see the predictability and efficiency of system development increase and it should see cost decrease.
Predictability should be obtained because system architecture supports planning, risk management, and other activities with system-wide scope:
- Planning is supported, because the knowledge about the system's architecture enables a completeness check of the work breakdown for the development work and the identification of dependencies between work packages. It is also supported because the order of integration and the needed verification can be planned and optimized according to knowledge about the system architecture.
- Risk management is supported, e.g. because the system architecture determines the contribution of subsystems to the performance of the system-of-interest and thus needs to be known for quantifying the influence of risks in subsystem development on the overall risk profile of system development. This is applicable for both project risk (the risk of project failure or deviation from its objectives) and product risk (the risk of undesirably or even dangerously wrong product performance).
- Staying in control of cyber-security will only be possible if the system-of-interest as a whole is architected for security, because intruders will find its weakest spot to attack it. Since the weakest spot may be an interface or a certain interaction pattern of several components, it is important to consider the system as a whole when foreseeing security measures and allocating them to the system's entities.
- Technologies based on artificial intelligence (AI) will require the validation of data used for training the AI, but maybe also the continuous collection of data about the performance of the AI itself. This requires governance across the system about the acquisition and aggregation of data to stay in control of data validity and availability. Control is also necessary about the impact of the...
