Application-level Time-Out Support for Real-Time Embedded Systems

Abstract A common requirement for many applications and services like, for example, membership protocols, is the availability of a class of functions for managing timeouts, i.e., objects that schedule an event, typically a function call, to be generated after a given amount of time. This paper describes an application-level time-out management systemthat exploits multiple alarm execution threads in order to reduce alarm execution congestion and consequent run-time violations. Through some experimental results it is shown howand in which cases this system can be made able to fulfill realtime requirements in spiteof this congestion.