2. Schedulability tests

2.1 Principles

When the scheduling policy is given, for example, it may be constrained by the real-time kernel used, the question asked is: is the system schedulable with the chosen policy? The test used is then called a schedulability test. It is used to demonstrate that, in the worst possible case corresponding to the system, all time constraints are satisfied. This means ensuring that the test applies to the chosen application model, that the worst case can be characterized, and that the test is viable, i.e. that the worst case taken into account actually corresponds to a worst case in terms of scheduling. An important factor is the time complexity of the test: to scale up to large applications (several hundred tasks), it is preferable for the test complexity to be polynomial,...