Software-hardware interdependent reliability assessment technique for software-intensive complex systems

Abstract

The current reliability testing methods cannot precisely evaluate the comprehensive reliability of SoftwareIntensive Complex Systems. In order to solve the problem, it is necessary to research the method of carrying out the softwarehardware interdependent reliability test. In this thesis, we do research on Software-hardware interdependent fault mode of Software-Intensive Complex Systems by collecting and studying the faults occurred in practical filed use. Then we design the software-hardware testing profile by adding software test profile to the reliability test profile, and put forward the idea about designing the test environment of software-hardware testing, which constitute a whole plan of software-hardware testing. And then a case study is demonstrated, which testifies the validation of the technique.Software-Intensive Complex Systems are those whose software plays a major role in the completion of required function or task, or in the development, operation, or the evolution of complex systems. The functions of complex software-intensive systems such as industrial control, avionics, satellite, telecommunications and other complex computer systems with hardware subsystems and software subsystems, can only be realized through the help of the embedded computer system featured by hardware and software together. With the extensive application of computer technology industries, complex software-intensive systems playing an increasingly important role in industrial products and people's daily lives are more closely related than ever.The operation mode of the Software-Intensive Complex Systems are becoming more complex more powerful and used under different operation environments, which calls for a higher demand for their reliability. In the mean time, the status of the software is becoming increasingly important and the soul of the system playing a key role in the realization of digital, intelligence, network and integration. Traditional reliability test becomes useless because it doesn't consider running the software operating profile and fails to test the function of the software adequately and on the other hand, when it comes to software test , we can't consider the impact of the hardware environment on the software functions, which makes it very difficult to detect how hardware environment stimulates the software-related fault of the system. Therefore, in independent hardware reliability testing and software testing exist loopholes and shortcomings, followed by:(1) The interdependent effect of the hardware reliability testing data and software testing data is poor, which can not be used for comprehensive reliability evaluation.(2) When reliability qualification test is performed, we are focusing on the performance under a variety of environmental stresses. Although when we test the hardware, we actually test the software at the same time, however, the test featuring a very small coverage of software functions, neglecting the faults that happen under a certain stress and a software is operating. For example, in...

Survey of combined hardware–software reliability prediction approaches from architectural and system failure viewpoint

“…However, literature addressing reliability evaluation of combined hardwaresoftware is limited in number (Park et al 2012;Hayakawa et al 2001). The existing literature can be broadly classified into two categories: (a) reliability estimation using failure data (time domain/frequency domain) at testing phase of the system development (Ji et al 2014;Park et al 2012;Welke et al 1995) and (b) reliability prediction using analytical approaches (architecture/component based) at the early stages of the system development (Tumer and Smidts 2011;Friedman et al 1992;Teng et al 2006).…”

“…A group of researchers considered that hardware and software failures are independent of each other. They did not explicitly model interaction failures (Ji et al 2014;Laprie and Kanoun 1992;Welke et al 1995;Vemuri and Dugan 1999). For example, Ji et al (2014) used operational profile in designing software-hardware test profiles but ignored HW-SW interactions.…”

“…They did not explicitly model interaction failures (Ji et al 2014;Laprie and Kanoun 1992;Welke et al 1995;Vemuri and Dugan 1999). For example, Ji et al (2014) used operational profile in designing software-hardware test profiles but ignored HW-SW interactions. They identified different hardware and software failure modes but did not include failure modes pertaining to interaction failure.…”

“…However, successive research in this domain has identified system may not fail only due to hardware specific or software specific failures. Failures attributable to improper interactions between hardware-software components also play a major role in system malfunctioning (Iyer and Velardi 1985;Ji et al 2014;Laprie and Kanoun 1992;Sumita and Masuda 1986). This indicates that reliability should be studied from the perspective of the combined hardware-software system.…”

“…All these reviews do not distinguish reliability estimation (Ji et al 2014;Park et al 2012;Welke et al 1995) from reliability prediction approaches (Tumer and Smidts 2011; Friedman et al 1992;Teng et al 2006). However, we restrain the scope of our study to reliability prediction in the early phases of development.…”