Book Image

The Complete Edition - Software Engineering for Real-Time Systems

By : Jim Cooling
Book Image

The Complete Edition - Software Engineering for Real-Time Systems

By: Jim Cooling

Overview of this book

From air traffic control systems to network multimedia systems, real-time systems are everywhere. The correctness of the real-time system depends on the physical instant and the logical results of the computations. This book provides an elaborate introduction to software engineering for real-time systems, including a range of activities and methods required to produce a great real-time system. The book kicks off by describing real-time systems, their applications, and their impact on software design. You will learn the concepts of software and program design, as well as the different types of programming, software errors, and software life cycles, and how a multitasking structure benefits a system design. Moving ahead, you will learn why diagrams and diagramming plays a critical role in the software development process. You will practice documenting code-related work using Unified Modeling Language (UML), and analyze and test source code in both host and target systems to understand why performance is a key design-driver in applications. Next, you will develop a design strategy to overcome critical and fault-tolerant systems, and learn the importance of documentation in system design. By the end of this book, you will have sound knowledge and skills for developing real-time embedded systems.
Table of Contents (16 chapters)
Preface
15
Glossary of terms

13.5 Middle-Out (Risk-Driven) Performance Modeling

As pointed out earlier, performance modeling is all about risk reduction in general. Now, it isn't all that unusual during development to find certain areas that are critical to the design. They are ones that, should they fail, will bring the project crashing down. In areas such as these, we need great confidence (backed by evidence) that the design will (or should) meet its performance requirements.

If our time-budgeting estimates can be supported with believable data (such as information obtained from earlier projects), then we are on solid ground. If not, however, the project is exposed to tremendous risk of failure. So, in circumstances like this, what can be done to mitigate this risk? One solution is to take a middle-out design approach and apply it to the critical items. In such cases, specific, localized modeling is done to provide confidence in the proposed design. Moreover, it must be done quite early in the design...