Mature research advances in scheduling theory show that carefully-crafted concurrent computational models permit static analysis of real-time behavior. This evidence enables designers to consider using suitable forms of explicit concurrency to model the inherent concurrency of real-time systems. The Ravenscar Profile, a specifically tailored subset of the Ada 95 tasking model, defines a compact and efficient concurrent computational model, especially suited for the development of high integrity, high efficiency real-time systems. Ravenscar runtimes can be implemented by small, efficient, reliable and certifiable kernels. At least two such implementations already exist and are being industrially deployed. The simplicity and intrinsic determinism of Ravenscar kernels facilitate the definition of metrics that cater for very accurate characterization of the dynamic behavior of the runtime and of the execution time of its primitives. Accurate runtime metrics enable forms of response time analysis that minimize the pessimism in the prediction of the runtime influence on the application. This is especially useful for concurrent systems that exhibit significant dependency on runtime support services. This paper recalls the motivations of the Ravenscar Profile, outlines the definition of it and formulates a precise characterisation of the associated runtime metrics.
On the Dynamic Semantics and the Timing Behaviour of Ravenscar Kernels
VARDANEGA, TULLIO;
2005
Abstract
Mature research advances in scheduling theory show that carefully-crafted concurrent computational models permit static analysis of real-time behavior. This evidence enables designers to consider using suitable forms of explicit concurrency to model the inherent concurrency of real-time systems. The Ravenscar Profile, a specifically tailored subset of the Ada 95 tasking model, defines a compact and efficient concurrent computational model, especially suited for the development of high integrity, high efficiency real-time systems. Ravenscar runtimes can be implemented by small, efficient, reliable and certifiable kernels. At least two such implementations already exist and are being industrially deployed. The simplicity and intrinsic determinism of Ravenscar kernels facilitate the definition of metrics that cater for very accurate characterization of the dynamic behavior of the runtime and of the execution time of its primitives. Accurate runtime metrics enable forms of response time analysis that minimize the pessimism in the prediction of the runtime influence on the application. This is especially useful for concurrent systems that exhibit significant dependency on runtime support services. This paper recalls the motivations of the Ravenscar Profile, outlines the definition of it and formulates a precise characterisation of the associated runtime metrics.Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.