Infact, independent applications with critical timing constraints sometimes fail to be schedulable by the standard Linux scheduler. Clearly, this happens when using the default Linux priority for those applications, as we all know how easy is to make a multimedia player show a “glitch” while playing an audio sample. Unfortunately, this happens also when using the special “real-time” scheduling policies, because, even though the scheduler gives priority to those applications w.r.t. the other Linux tasks, the Round-Robin or FIFO schedulers built into the standard Linux scheduler are inappropriate and introduce timing dependencies among RT tasks themselves.
On the other hand, when using the AQuoSA scheduler, it is possible to provide temporal isolation among those applications, and between those applications and the other non-RT tasks in the system, so that the Operating System manages to succesfully guarantee timing constraints to all of them.
A simple application scenario highlighting and
clarifying the concept is shown on the experiment
page of the AQuoSA website.”
Link: aquosa.sourceforge.net
Category:
- Linux
