A new technique has been developed for organizing (or reorganizing) system design to enhance fault diagnosability. This technique is called the controllability, observability, and maintenance engineering technique, or COMET. Using graph-theoretical analysis, one can systematically apply COMET to a proposed or an existing digital system to determine the placement of control, access, and monitor points for diagnostic testing. In addition, it provides a means of studying the trade-offs between fault resolvability and the cost of maintenance hardware and/or packaging. COMET offers an orderly approach to implementing an overall diagnostic design by providing guidelines in early design stages. A design developed using COMET has the following advantages: trouble location manual data can be generated without the use of fault simulation, multiple faults and/or nonclassical faults are locatable if they are detectable, and diagnostic or trouble-location information can be easily updated in accordance with hardware changes. Studies indicate that applying COMET to an existing processor design would require a modest increase in hardware of less than 10 percent.
[1]
C. V. Ramamoorthy,et al.
A structural theory of machine diagnosis
,
1899,
AFIPS '67 (Spring).
[2]
Herbert Y. Chang,et al.
The Design of a Microprogrammed Self-Checking Processor of an Electronic Switching System
,
1973,
IEEE Transactions on Computers.
[3]
Herbert Y. Chang,et al.
Maintenance Techniques of a Microprogrammed Self-Checking Control Complex of an Electronic Switching System
,
1973,
IEEE Transactions on Computers.
[4]
CHITOOR V. RAMAMOORTHY,et al.
System Segmentation for the Parallel Diagnosis of Computers
,
1971,
IEEE Transactions on Computers.