Distributed real-time task monitoring in the safety-critical system Melody

Safety-critical systems typically operate in unpredictable environments. Requirements for safety and reliability are in conflict with those for real-time responsiveness. Due to unpredictable environmental needs there is no static trade-off between measures to accommodate the conflicting objectives. Instead every feature or operating system service has to be adaptive. The distributed operating system MELODY was developed for supporting safety-critical applications under such constraints. In earlier phases real-time monitoring features like file monitoring and integration control had been implemented. This paper focuses on distributed task monitoring. In order to as early as possible abort task instances which cannot meet their deadline sub-deadlines for their location, allocation, locking phases are derived. These are used for creating three different monitoring models in which task instances would be aborted whenever they are about to miss the sub-deadline corresponding to the model. Extensive distributed experiments are reported in which a variety of refined models are compared to a Base model. Despite the considerable overhead for task monitoring the results of employing the distributed task monitors are very favorable, with respect to minimizing the total number of deadline failures as well as regarding survivability.