Q 326. Fault Injection Testing is a process where faults are deliberately injected during system testing. What are the benefits of such a testing approach and highlight its application areas.

 

 

Note for website visitors - Two questions are asked every week on this platform. One on Tuesday and the other on Friday.

• All questions so far can be seen here - https://www.benchmarksixsigma.com/forum/lean-six-sigma-business-excellence-questions/
• Please visit the forum home page at https://www.benchmarksixsigma.com/forum/ to respond to the latest question open till the next Tuesday/ Friday evening 5 PM as per Indian Standard Time. Questions launched on Tuesdays are open till Friday and questions launched on Friday are open till Tuesday. 
• The best answer is always shown at the top among responses and the author finds honorable mention in our Business Excellence dictionary at https://www.benchmarksixsigma.com/forum/business-excellence-dictionary-glossary/ along with the related term

Fault Injection: Overview
In Fault Injection experiments, various faults are injected into a simulation model of the target system or a hardware-and-software prototype of the system. The behavior of the system in the proximity of each fault is then observed and classified. Parameters that can be considered based on such experiments include the probability that a fault will create an error and the probability that the system will successfully perform the actions required to recover from that error. 

 

These actions consist of recognizing the fault, identifying the system component influenced by the fault, and taking appropriate recovery action, involving system reconfiguration. Each of these actions necessitates time that is not continuous but may change from one fault to another and depend on the overall workload. Thus, fault injection experiments and presenting estimates for the coverage factor can also estimate the individual delay's distribution associated with each of the above actions.

 

Also, fault injection experiments can be used to evaluate and validate the system dependability. For instance, errors in the implementation of fault tolerance mechanisms can be identified. System components whose negligence is further likely to result in a total system crash can be identified. Also, the effect of the system's workload on the dependability can be witnessed.

 

Fault Injection: Application
Fault injection must be applied to measure the coverage and latency parameters, study error propagation, and analyze the relationship between the system's workload and its fault handling capabilities. Another exciting utilization of fault injection systems is to evaluate the effect of transient faults on the availability of highly reliable systems. These systems were capable of improving from the transient faults but still had misused time doing that, thus diminishing the availability.

 

Various fault injectors have been acquired and are currently in use. Studies comparing several fault injectors have been administered, concluding that two fault injectors may either endorse or complement each other. The latter occurs if they satisfy different faults. The different strategies to fault injection result in quite other characteristics of the corresponding tools. Some of these differences are reviewed in the below table which compares the properties of four approaches to fault injection. 

 

[Image Source: Fault-Tolerant Systems By Israel Koren and C. Mani Krishna]

 

All fault injection schemes expect a well-defined fault prototype, which should describe as closely as possible the faults that one requires to see during the endurance of the target system. A fault model must describe the types of defects, their location & duration, and, possibly, the statistical correlations of these properties. The fault models used in currently available fault injection tools deviate considerably, from very detailed device-level faults (for example, a delay fault on a distinct wire) to simplified functional level faults (such as an erroneous adder output).

Sourabh Nandi has provided the best answer to this question.