A clear, rigorous look at real-time scheduling that balances guarantees with practical performance.
This book presents an online scheduling algorithm called DD* for real-time systems. It explains how the method combines a first-deadline-first approach with adaptive timing, to perform well both when the system is underloaded and when it is overloaded. The work shows how to measure performance against a clairvoyant scheduler and proves optimality in underloaded conditions.
Readers will see how the algorithm uses three data structures to manage tasks and deadlines, how it handles different kinds of events, and why a competitive factor of one-fourth can be achieved. The text also covers the theoretical limits of online scheduling and offers a detailed, step-by-step description of the method and its guarantees.
Ideal for readers of research in real-time systems and practitioners seeking solid, provable scheduling strategies.
"synopsis" may belong to another edition of this title.