Implementing uniform treatment of hardware and software faults with distributed queues
Phillip Louis Amicucci · Montana State University ScholarWorks (Montana State University) · 1993
This thesis explores the use of replicated queues with distributed execution of recovery blocks (DRB) as an approach for uniform treatment of hardware and software faults.An approach to incorporating the capabilities of DRB's into a loadsharing multiprocessing scheme using replicated queues is discussed.A simulation aimed at testing the execution efficiency and fault tolerance of the scheme in a realtime environment has been conducted using IBM's OS/2 2.0's operating system.The simulated completely connected network was developed and successfully tested on a single 386DX/33 PC.The results indicate the feasibility of achieving tolerance to hardware and software faults in a real-time computer system.The effective use of the schemes distributed execution of recovery blocks with replicated queues deserves serious consideration.