Formal description of realtime systems: a review

Abstract

Work on the formal description and analysis of realtime systems has followed two paths. On one side, it has considered the specification of realtime systems, the design of language constructs for realtime programming, and semantic models to describe the properties of realtime programs. On the other side, there has been a large body of work analysing the performance of realtime systems in terms of scheduling theory, relating program modules with their use of resources, and determining timing characteristics, especially in relation to hard realtime problems. The paper reviews some of this work and examines how realtime computations are modelled in a proof-theoretic framework and in scheduling analyses. The object of this review, and indeed of other contemporary work in the field, is to examine whether it is possible to relate issues of correctness and performance, e.g., interpreting the semantics of execution of realtime programs in terms of scheduling solutions to realtime problems.