At the core of slim mobile phones, or complex information systems for airplanes, are real-time kernels that schedule tightly time-constrained parallel processes. To avoid the maintenance nightmare of real-time systems designed in an ad hoc manner, the US Office of Naval Research launched a real-time systems initiative in the 1980s. This coordinated effort led to a wealth of new scheduling approaches, protocols, and standards that greatly improved the state of the art. These found their way into high-visibility projects, such as a major software update for the global positioning system (GPS) constellation, and the design of the Space Station.

This paper offers a comprehensive review of the developments brought about by decades of research in the field of real-time systems. The first sections look at the pioneering work of Liu and Layland in fixed priority scheduling and, moving beyond this important, yet rather limiting framework, dynamic prio!rity scheduling, as implemented, for instance, in the earliest deadline first algorithm. When time constraints are somewhat loose, up to the point where some tasks can even be dropped (for example, in video broadcasting), soft real-time systems offer better bang for the buck. To deal with even more complex systems, control theory-based approaches using feedback models can be helpful. Yet, despite all of these developments, challenges remain, particularly within large cooperating systems.

This comprehensive and easy-to-read paper, together with its extensive survey of the literature, should be required reading for anyone interested in making a contribution to the field of real-time systems.