Browse Books

Reviewer: Maulik A Dave

In the field of operating systems (OSs), a real-time task differs from other types of tasks because it has a deadline. If a hard task's deadline is not met, this can cause a catastrophe. This book describes and analyzes many algorithms for scheduling such tasks in real-time systems. The book also introduces existing real-time systems and real-time system standards, and discusses some system kernel design considerations. Chapter 1 introduces three types of real-time system tasks: hard, firm, and soft. It explains general real-time systems with examples and by comparing their features with those of time-sharing systems. It discusses desirable features of real-time systems and introduces ways to achieve them. Chapter 2 defines a general scheduling problem and related terms such as task, precedence constraints, resource constraints, and timing constraints. It introduces classifications of scheduling algorithms and performance metrics for scheduling. The chapter ends by describing scheduling anomalies, which are those situations where schedule length increases with more computational power, or with weaker constraints. Chapter 3 describes various algorithms for aperiodic task scheduling, starting with the earliest due date (EDD) algorithm for scheduling aperiodic tasks on a single processor, minimizing maximum lateness. Next, the earliest deadline first (EDF) algorithm solves the same problem, but allows for preemption and dynamic arrivals. The chapter ends with algorithms for nonpreemptive scheduling and scheduling with precedence constraints. Chapter 4, on periodic task scheduling, describes four algorithms, namely, timeline scheduling, rate monotonic, EDF, and deadline monotonic, and presents parameters used on periodic tasks. Chapter 5 describes algorithms for scheduling tasks that contain both hard periodic tasks with fixed priorities and soft aperiodic tasks. The algorithms include background scheduling, polling server, deferrable server, priority exchange, sporadic server, and slack stealing. Chapter 6 describes algorithms for scheduling tasks containing periodic and/or aperiodic tasks with dynamic priorities. The algorithms include dynamic priority exchange server, dynamic sporadic server, total bandwidth server, earliest deadline late server, and constant bandwidth server. Some tasks may share resources. Chapter 7 discusses the resource access protocols for such tasks, including nonpreemptive protocols, highest locker priority, priority inheritance protocol, priority ceiling protocol, and stack resource policy. The description of each protocol includes how to compute blocking time, properties of the protocol, and implementation considerations. The preemptive tasks introduce a preemption cost into the scheduling. Chapter 8 discusses techniques to limit preemption costs. It describes techniques such as nonpreemptive scheduling, preemption thresholds, deferred preemption, task splitting, and selecting preemption costs. Each technique is explained along with a feasibility analysis of the technique. The chapter ends with a comparative analysis of the techniques. Chapter 9 is on handling overloads. When the computation time demanded by a task set exceeds the available processing time, the system is said to be overloaded. The chapter discusses handling overloads caused by both periodic and aperiodic tasks in detail. Chapter 10 discusses design and implementation considerations for the kernel of a real-time system. It describes the general structure, data structures, process states, time management, scheduling algorithms, scheduling mechanisms, task management, and inter-task communication mechanisms in the kernel of a small educational real-time system. Chapter 11 introduces control applications, such as peg-in-hole insertion, surface cleaning, object tactile exploration, and catching moving objects as examples of real-time robotics applications. It includes the source code of a robot control using real-time primitives. Chapter 12 gives an overview of some existing real-time systems. The chapter begins with various standards of real-time systems, such as RT-POSIX, OSEK/VDX, ARINC-APEX, and Micro-ITRON. It introduces both commercial- and research-level real-time systems. The chapter ends with a discussion of some of the developer tools. A strong background in algorithms and OSs is required to understand this book, as is a strong background in mathematics. Each chapter ends with a set of exercises; the solutions are in the last chapter (13). The book can be used as undergraduate- and graduate-level textbook for courses involving real-time systems. The initial chapters are a good introduction to real-time systems for beginners. Professionals and researchers working with real-time systems will find the scheduling algorithms useful. Online Computing Reviews Service