| Abstract |
To deal with a large variety of workloads indifferent application domains in real-time embedded systems,a number of expressive task models have been developed. Foreach individual task model, researchers tend to develop differenttypes of techniques for deriving schedulability tests with differentcomputation complexity and performance. In this paper, wepresent a general schedulability analysis framework, namely thek2U framework, that can be potentially applied to analyzea large set of real-time task models under any fixed-priorityscheduling algorithm, on both uniprocessor and multiprocessorscheduling. The key to k2U is a k-point effective schedulabilitytest, which can be viewed as a ?blackbox? interface. For anytask model, if a corresponding k-point effective schedulabilitytest can be constructed, then a sufficient utilization-based testcan be automatically derived. We show the generality of k2U byapplying it to different task models, which results in new andimproved tests compared to the state-of-the-art.
|
