Tutorial on real-time scheduling - Grolleau - 2 col

Tutorial on real-time scheduling Emmanuel Grolleau Laboratoire d’Informatique Scientifique et Industrielle grolleau@ensma.fr another one has a period of 5 ms, and a third one is Abstract thtriggered by an interrupt, while a 4 one should send a command to an actuator every 20 ms. This presentation is a tutorial given as a survey of the There are two ways to program a (asynchronous, we basic problems arising in real-time schedulability are now always in the asynchronous hypothesis) real-analysis on uniprocessor systems. It mainly focuses on time system: event-driven programming, and time-driven on-line scheduling policies (fixed priority policies - FPP, programming. On one hand, in event-driven systems, and dynamic priority scheduling) on a preemptive events are interrupts. Events are either coming from scheduling scheme, and insists on the basic concepts of sensors (including a keyboard or a pointing device) or busy period and processor demand. This tutorial is an from the internal clock. Thus, a task is released by an extract of a lecture given in Masters of Computer event, treats it, and then waits for the next event. On the Science and most of the results presented here can be other hand, time-driven systems are based on a time found in books [Liu00][But04]. quantum: a task is awaken at its time, does a treatment anyway (even if nothing happened since its last release) 1. Introduction and then sleeps until its next release time. Active sensors can ...