System Architecture: Strategy and Product Development for Complex Systems

Citation

Crawley, E., Cameron, B., & Selva, D. (2015). System Architecture: Strategy and Product Development for Complex Systems. Pearson.

Chapter Summary

The book has 16 chapters in four parts, with appendices, chapter problems and an index. Pearson describes system architecture as the study of early decision making in complex systems: how to capture experience and analysis about early decisions, and how to choose architectures that meet stakeholder needs, integrate easily and evolve flexibly. The publisher adds that case studies by practitioners, from hybrid cars to communications networks to aircraft, show the science and art of the subject. The four parts move from ways of thinking, to analysing an existing architecture, to creating one, and finally to treating architecture as a set of decisions.

Part 1, Chapters 1-3: System Thinking

Chapter 1, "Introduction to System Architecture", covers the architecture of complex systems, the advantages of good architecture, the learning objectives and the organisation of the text. Chapter 2, "System Thinking", introduces systems and emergence and then works through four tasks: identify the system, its form and function; identify its entities and their form and function; identify the relationships among the entities; and examine emergence. Chapter 3, "Thinking about Complex Systems", covers complexity, the decomposition of systems, special logical relationships and ways of reasoning through complex systems, and ends with the architecture representation tools SysML and OPM.

Part 2, Chapters 4-8: Analysis of System Architecture

Chapter 4, "Form", looks at form in architecture, the analysis of form and of formal relationships, formal context, and form in software systems. Chapter 5, "Function", treats external function and value, internal function, functional interactions and functional architecture, and secondary value-related functions. Chapter 6, "System Architecture", puts form and function together and adds non-idealities, supporting layers and interfaces, operational behavior and reasoning with representations. Chapter 7, "Solution-Neutral Function and Concepts", moves from identifying a solution-neutral function to concepts, integrated concepts, and concepts of operations and services. Chapter 8, "From Concept to Architecture", develops a Level 1 and a Level 2 architecture, illustrated with a home data network, and shows how to modularize the system.

Part 3, Chapters 9-13: Creating System Architecture

Chapter 9, "The Role of the Architect", deals with ambiguity and the product development process. Chapter 10, "Upstream and Downstream Influences on System Architecture", lists upstream influences (corporate strategy, marketing, regulation and technology infusion) and downstream ones (implementation, operations, design for X, and product and system evolution), ending with the Architecture Business Case Decision. Chapter 11, "Translating Needs into Goals", covers identifying beneficiaries and stakeholders, characterizing needs, interpreting them as goals and prioritizing the goals. Chapter 12, "Applying Creativity to Generating a Concept", works from creativity to developing, expanding, refining and selecting concepts. Chapter 13, "Decomposition as a Tool for Managing Complexity", takes up understanding and managing complexity.

Part 4, Chapters 14-16: Architecture as Decisions

Chapter 14, "System Architecture as a Decision-Making Process", opens with the formulation of the Apollo architecture decision problem and goes on to decision support tools. Chapter 15, "Reasoning about Architectural Tradespaces", covers tradespace basics, the Pareto frontier, the structure of a tradespace, sensitivity analysis and the organization of architectural decisions. Chapter 16, "Formulating and Solving System Architecture Optimization Problems", formulates such problems, works an example (the NEOSS Earth-observing satellite system for NASA), looks at patterns in architecting decisions and at large-scale problems, and then at methods for solving them.

Key Concepts

Part 1: System thinking

• Emergence: behavior of a system that arises from the relationships among its entities rather than from any one of them.
• Form and function: the two views, what a system is and what it does, from which the book's analysis of architecture is built.
• Decomposition and complexity: how a system is divided into parts, and how that division manages or hides complexity.
• SysML and OPM: two modeling languages for representing architecture.

Part 2: Analysis

• External function and value: what a system does for those outside it, and the value this creates.
• Solution-neutral function and concept: a statement of what must be done that does not fix how, and the concept that proposes how.
• Integrated concepts and concepts of operations: the combination of concepts into a whole, and the description of how the system is used.
• Modularization: grouping elements of the system into modules at successive levels of architecture.

Parts 3-4: Creation and decision

• Stakeholders, needs and goals: the chain from the people affected, through their needs, to prioritized goals.
• Upstream and downstream influences: strategy, marketing, regulation and technology before the architecture; implementation, operations, design for X and evolution after it.
• Tradespace and Pareto frontier: the set of possible architectures plotted by their attributes, and the subset not dominated by others.
• Sensitivity analysis: testing how conclusions change as assumptions change.
• Architecture optimization: formulating the choice of architecture as an optimization problem and solving it.

Critical Analysis

Strengths

• A vocabulary built from first principles. Chapters 2 to 8 build form, function, concept and architecture in a stated sequence, with four explicit tasks in Chapter 2, so the reader acquires a consistent way of describing any system before being asked to design one.
• Both the human and the quantitative side. Part 3 deals with the architect's role, strategy, marketing, regulation and stakeholder needs, while Part 4 turns to decision support, tradespaces and optimization.
• Worked cases from practice. The contents list names a home data network, the Apollo architecture decision and the NEOSS Earth-observing satellite system as worked examples, and Pearson highlights case studies written by leading practitioners, from hybrid cars to communications networks to aircraft.
• Teaching apparatus. Most chapters open with an introduction and close with a summary and references, and the book includes a chapter problems section.

Limitations

• Specialized scope. Pearson presents the text for courses in engineering and technical management. The book is about early architectural decisions, not about requirements management, detailed design, verification or project management, which appear only as upstream and downstream influences in Chapter 10.
• A demanding last part. Chapters 14 to 16 turn to decision theory, Pareto analysis and formal optimization, so they draw on quantitative methods (Pareto analysis, sensitivity analysis, optimization) that the headings of the earlier chapters do not show.
• Representation tools in one section. SysML and OPM appear in a single section of Chapter 3 (Section 3.6), so readers who want to learn a modeling language will need a separate source.
• A long path to the creative chapters. The chapters on generating concepts (Chapter 12) and the architect's role (Chapter 9) come after eight chapters of analysis, which suits a course but may feel slow for a practitioner who wants to get straight to practice.

Where It Fits in the Course

This is the architecture-focused text in the systems engineering list, best read after a general introduction. Holt's Systems Engineering Demystified is a practical first introduction that uses SysML diagrams, and Dym, Little and Orwin's Engineering Design: A Project-Based Introduction is a broad introduction to design methodology and problem-solving. Kossiakoff and colleagues' Systems Engineering Principles and Practice and Blanchard and Fabrycky's Systems Engineering and Analysis both follow the life cycle of a system and supply the processes and tools that frame where architecture fits.

Buede and Miller's The Engineering Design of Systems is the closest neighbor: it is a model-based treatment that goes from stakeholder needs and requirements through functional and physical architecture to integration and qualification. Crawley, Cameron and Selva go deeper into form, function and concept, and into architecture as a decision problem with tradespaces and optimization, while Buede and Miller give a wider sweep of the engineering design process. The two are complementary, and a graduate course can use one for the process and the other for the architecture.

How to Use This Book

The book is for graduate students and for engineers in technical management who have worked on development projects and know basic systems engineering ideas such as requirements and decomposition. Pearson presents it for courses in engineering and technical management.

A practical plan:

• Core: Chapters 2, 4, 5 and 6, which build the form and function vocabulary used everywhere else, followed by Chapters 7 and 8 on concept and architecture.
• For a project course: Chapters 9 to 12 on the architect's role, influences, stakeholder needs and concept generation.
• For quantitative readers: Chapters 14 to 16 on decisions, tradespaces and optimization; work through the NEOSS example of Chapter 16 from formulation to solution.
• Alongside: pick a real product and apply the four tasks of Chapter 2 and the form and function analysis to it before reading further, and use the chapter problems to check each step.