Foreword

Real-time embedded systems have become an integral part of our technological and social space. But is the engineering profession equipped with the right knowledge to develop those systems in ways dictated by the economic and safety aspects? Likely yes. But the knowledge is fragmented and scattered among different engineering disciplines and computing sciences. Seldom anyone of us has the clear picture of the whole. If so, then parts of it are at an abstract level. That poses a question whether the academic system provides education in a way holistic enough to prepare graduates to embark on the development of real-time embedded systems, frequently complex and imposing safety requirements. How many electrical and computer engineering departments offer subjects focusing on the application-dependent specialized communication networks used to connect embedded nodes in distributed real-time systems. If so, then the discussion is confined to the Controller Area Network (CAN), or sometimes FlexRay, in the context of automotive applications—usually a small unit of an embedded systems subject. (The impression might be that specialized communication networks are mostly used in automotive applications.) The requirement for the underlying network technology to provide real-time guarantees for message transmissions is central to proper functioning of real-time systems. Most of computer engineering streams teach operating systems. But real-time aspects are scantly covered. Computer science students, on the other hand, have very little, if any, exposure to the “physicality” of the real systems the real-time operating systems are intended to interact with. Does this put computer science graduates in a disadvantaged position? In the late 1990s and early 2000s, I was involved in the Sun Microsystems lead initiative to develop real-time extensions for the Java language. The working group comprised professionals mostly from industry with backgrounds largely in computing sciences. I was taken aback by the slow pace of the process. On reflection, the lack of exposure to the actual real-time systems in different application areas and their physicality was likely to be behind difficulties to identify generic functional requirements to be implemented by the intended extensions.

In the second part of 1980s, I was teaching digital control to the final year students of the electrical engineering course. The lab experiments to illustrate different control algorithms were designed around the, at that time, already antiquated Data General microNOVA MP/200 minicomputer, running one of the few real-time operating systems commercially available at that time—QNX, if I remember correctly. Showing things work was fun. But students’ insight into working of the whole system stopped at the system level commands of the operating systems. The mystery had to be revealed by discussing hypothetical implementations of the system level calls and interaction with the operating system kernel—of course at the expense of the digital control subject. At that time, seldom any electrical engineering curriculum had a separate subject dedicated to operating systems. Of frustration and to avoid the “black box” approach to illustrating control systems in action, I have written in C a simple multitasking real-time executive for MS-DOS-based platforms, to be run on an IBM PC (Intel 8088). Students were provided with the implementation documentation in addition to theoretical background; quite a lot of pages to study. But the reward was substantial: they were now in full “control.” With the support of an enterprising post-graduate student, the executive was intended to be grown into more robust RTOS with a view for commercialization. But it was never to be. Academic life has other priorities. Around 1992, I decided to harness the MINIX operating system, which I then taught to the final-year graduate students, to run my real-time control lab experiments to illustrate control algorithms in their supporting real-time operating system environment. But soon after that came the Linux kernel.

If you are one of those professionals with the compartmented knowledge, particularly with the electrical and computer engineering or software engineering background, with not much theoretical knowledge of and practical exposure to real-time operating systems, this book is certainly an invaluable help to “close the loop” in your knowledge, and to develop an insight into how things work in the realm of real-time systems. Readers with a background in computer science will benefit from the hands-on approach, and a comprehensive overview of the aspects of control theory and signal processing relevant to the real-time systems. The book also discusses a range of advanced topics which will allow computer science professionals to stay up-to-date with the recent developments and emerging trends.

The book was written by two Italian researchers from the Italian National Research Council (CNR) actively working in the area of real-time (embedded) operating systems, with a considerable background in control and communication systems, and a history of the development of actual real-time systems. Both authors are also involved in teaching several courses related to these topics at Politecnico di Torino and University of Padova.

The book has been written with a remarkable clarity, which is particularly appreciated whilst reading the section on real-time scheduling analysis. The presentation of real-time scheduling is probably the best in terms of clarity I have ever read in the professional literature. Easy to understand, which is important for busy professionals keen to acquire (or refresh) new knowledge without being bogged down in a convoluted narrative and an excessive detail overload. The authors managed to largely avoid theoretical only presentation of the subject, which frequently affects books on operating systems. Selected concepts are illustrated by practical programming examples developed for the Linux and FreeRTOS operating systems. As the authors stated: Linux has a potential to evolve in a fully fledged real-time operating system; FreeRTOS, on the other hand, gives a taste of an operating system for small footprint applications typical of most of embedded systems. Irrespective of the rationale for this choice, the availability of the programming examples allows the reader to develop insight in to the generic implementation issues transferrable to other real-time (embedded) operating systems.

This book is an indispensable addition to the professional library of anyone who wishes to gain a thorough understanding of the real-time systems from the operating systems perspective, and to stay up to date with the recent trends and actual developments of the open-source real-time operating systems.

Richard Zurawski
ISA Group, San Francisco, California