Report of Session: Flexible Scheduling in Ada Chair: Alan Burns Rapporteurs: Andy Wellings, Tullio Vardanega 1 Introduction The goal of this session is to determine whether Ada s support for scheduling is adequate for modern and future real-time systems. The chair presents an overview of the Ada model of priority-based scheduling and argues that the ARM needs to be changed to allow other scheduling schemes to be de ned. He also notes that, although implementations are currently allowed to provide other schemes, experience has shown that this will not happen, unless the language de nition sanctions what other schemes should be supported. The chair s view is that adding new schemes would not imply that they should all be supported by an implementation of Annex D. If they were supported, however, they should be supported as speci ed in the Annex. Schemes that Ada could support include: new paradigms such as EDF, Value-based, Round Robin, Servers (of various types), Combined paradigms, as well as primitives for roll-your-own user-de ned scheduling. The notion of Combined scheduling would include: Within a partition 2 Alternative Scheduling Schemes Alan Burns then proposes a model that leaves the underlying Ada scheduling model
/lp/association-for-computing-machinery/report-of-session-flexible-scheduling-in-ada-3o4VhwLJAK