Review: Model-Based System Architecture

Summary: The book "Model-Based System Architecture" is not easy to read. It makes high demands on the reader, and even higher demands on itself. Hardly a statement is made without a relevant source. Accordingly, a clean foundation is laid first (formalized in SysML), which describes what system architecture actually is. Building on this, the guiding principles are presented. With this foundation, different topics are then dealt with chapter by chapter, whereby the chapters are quite independent and do not have to be consumed in the printed order. On the contrary, the topics give the impression of a colorful bouquet of flowers and deal with requirements analysis, viewpoints, stakeholders, processes, analyses and variants. The book consistently develops many of the authors' already familiar ideas, such as FAS and SYSMOD. A system for virtual museum tours runs through the book as an example.

Target group: Experienced systems engineers and practitioners who want to delve deeper into specific areas of model-based architecture.

Rating: ★★★★★ Working your way through this book is hard work, but it is rewarded by the fact that every sentence really counts.

Authors: Tim Weilkiens, Jesko G. Lamm, Stephan Roth, Markus Walker

Acquire: The book is available on paper (€135 Amazon Afiliate Link) and as PDF (€110), whereby here DRM from Adobe is used, which extremely limits its usefulness (cannot be used with Kindle or Linux, no printing possible). The high price is particularly unpleasant for students, but unfortunately not unusual for books of this caliber.

Content of Model-Based System Architecture

Details: Nothing in this book is taken for granted: this also applies to the definition of system architecture itself. In order to get a good one, various definitions are first analyzed. The result itself is then recorded as a SysML block diagram. The approach, for the definitions SysML is practiced in many places in the book. At first this was somewhat surprising, but it definitely has its charm and is very effective. Of course, this assumes that the reader understands SysML. There is a short 25-page introduction in the appendix.

Many definitions are documented in the form of SysML block diagrams.

Motivation is also not unimportant, MBSE to operate at all. Here, too, the reader is provided with solid arguments (Chapters 3, 5 and 6).

It then gets down to business Chapter 7in which patterns and principles are presented. Topics such as basic architecture, views of the model, heuristics and similar topics are addressed here.

Start with requirements

Requirements play an important role in identifying the right architecture. It is therefore understandable that a separate chapter is dedicated to this topic, "Requirements and Use Case Analysis" (Chapter 8). Later, the FAS method is introduced (Functional Architectures for Systems, Chapter 14), which starts where Chapter 8 stops. This is where a functional architecture is built on the basis of the use cases. This is one of the most exciting chapters in the book: functional architectures initially describe (surprise!) the functions of the system. However, these are initially technology-independent. This is extremely relevant today, in a world of rapid technological change. As an example of this, in Chapter 7 the example of gaining access to a vehicle. In the past, this was done with a mechanical key, today with a remote control built into the key, and tomorrow perhaps with an RFID tag. The function remains the same.

The chapters in between deal with several topics that are only loosely connected. Chapter 9 deals with viewpoints and perspectives on models. I think this is extremely valuable, because models are only useful if they are read. And to make a model readable, it must present relevant information in a way that appeals to the reader.

Stakeholders...

Stakeholders of the architecture are Chapter 10 dealt with. Ultimately, this is a list that will certainly sound familiar to experienced stakeholders. What is useful here is the tabular representation that shows how the relationships can be optimized ("win-win"). Chapter 11 then describes the roles, which logically builds on the previous chapter. This is more about topics such as responsibilities and tasks. This chapter provides valuable suggestions for engineers who want to develop in the direction of SE.

... processes...

Chapter 12processes, is rather thin. What is interesting here is the role that modeling plays in the processes, as these can be supported automatically. Also Chapter 13"Agile Approaches", doesn't have much to offer. I would describe it as the weakest chapter in the book. There is little more here than a brief introduction, a historical outline and further reading.

and variants

Chapter 14 in turn addresses a very important topic: variants. Various approaches such as FODA, CVL or OVM are touched on here. But most importantly, it shows how variant management can be practiced with SysML. This starts with a sensible package structure and uses a concrete example to show how to work with feature trees. What is shown here is extremely useful, and I would have liked more than just eight pages.

Chapter 14 describes practical variant management with SysML.

In Chapter 15 Here it is first clarified that this refers to enterprise architectures that are known in particular from the defense sector, such as NAF, DoDAF or TOGAF. I've come across such frameworks a few times and would have liked a 20-page introduction like this one.

The next chapters will focus on implementation. Chapter 16 addresses cross-cutting topics such as human-system interaction and risk. The short Chapter 17 deals with architecture evaluations (assessments) and helps to proceed in a more structured way instead of just using intuition. In Chapter 18 The topic is closely related to soft skills, which are taught in Chapter 19 be dealt with.

No book is complete without a look into the crystal ball. In Chapter 20 the authors ask whether the era of product lines is coming to an end and will be replaced by fully customized developments. They do not presume to answer this question, but ask many good thought-provoking questions.

Conclusion

The book Model-Based System Architecture is not for beginners! It requires the team to have a deep understanding of systems engineering and the basics of MBSE. But if this is given, then it is enormously valuable as a roadmap for better, cheaper and faster developed products.

Disclosure: This review is based on a free copy (eBook) from Wiley-Verlag.