A technique for monitoring a real-time multi-computer system under heavy loading

Abstract

Techniques are developed for the design of a monitor of a real-time multi-computer system that is under heavy loading. The first portion relates to the requirements of partitioning to aid in fault recognition and diagnostic routines. The dynamic reallocation of system time to the system tasks and fault monitoring in developed secondly. System reconfiguration of the partitioned subsystems restores the system to operation at a degraded level until faults are corrected. the paper discusses a Ship Combat Weapon System as an example of a large scale multi-computer system monitor.