|
Other FMEA
Sources
| |
Title: Markovian software safety measurement with reliability growth process
Author: Tokuno, K; Yamada, S
Abstract: This paper reconstructs a quantitative software safety/reliability model based on the Markovian software reliability one with imperfect debugging proposed by Yamada et al. [1], and provides a metrics of software safe ty defined as the probability that the system does not fall into hazardous states at a specified time point. Tokuno and Yamada [2] suggested two stochastic software safety assessment models assuming that the system may fall into unsafe states only when software failures occur. In contrast, the attention of this paper is directed to the event that the system causes hazardous conditions randomly in operation. In particular, we assume that some of debugging activities contribute to software safety improvement as well in software safety modeling. We refer to the difference of the definitions between software safety and reliability. Software safety is defined as the attribute that software systems do not induce unsafe conditions or states. Software systems in unsafe states lead to fatal accidents, mishaps, and hazards; for instance, financial losses or injuries to human life. By contrast, software reliability is defined as the attribute that systems can continue to operate according to the specifications without software failures; these are unacceptable departures from program operations caused by faults remaining in the systems. Accordingly, all of software failure-occurrences do not cause the problems relating to safety and systems may not always keep safe states even though they function in accordance with the specifications [3]. There are little techniques for assessing software safety, for example, Fault Tree Analysis (FTA) and Failure Mode and Effect Analysis (FMEA) are qualitative static methods. But recently, quantitative evaluation methods for measuring software safety in dynamic environment begin to be required since there are limitations oil analyzing the time-dependent state-transitions for safety-critical systems with FTA and FMEA. However, such methods scarcely exist. Several stochastic quantities for software safety/reliability measurement are derived from this model and numerical illustrations are also presented.
Source: PSAM 5: PROBABILISTIC SAFETY ASSESSMENT AND MANAGEMENT, VOLS 1-4
Publication Year: 2000
Volume:
Issue nr: 34
Pages: 2681 - 2686
Science Code: Computer Science, Artificial Intelligence; Engineering, Civil; Engineering, Electrical & Electronic; Operations Research & Management Science; Nuclear Science & Technology
Tags:
|