Software Reliability Growth Models Exposing Software Field Failures Test Time Test Time Software Reliability Objective Test Time Figure 1: Example Srgm

Content Provider	Semantic Scholar
Author	Siok, Michael F. Whittaker, C. J.
Abstract	G rowing the quality of software during a fighter aircraft avionics software development project is a natural outcome of conducting a disciplined software development effort. Through each step of the process, the software product evolves toward its end-product state, implementing more capabilities as it progresses toward its scheduled release date. Through a comprehensive and directed test regimen, the risk of releasing serious defects in the delivered software diminishes greatly as testing progresses, demonstrating software reliability growth. Once software development is complete and aircraft are being delivered, on-site operations teams manage the day-today operations of the deployed aircraft. Any malfunctioning avionic computers (e.g., computer will not boot, improper return to operation following a reboot, selected avion-ics operations in obvious violation of functional specification) may be managed through maintenance event actions by either replacing faulty equipment or reprogramming the computer and verifying a return to normal operation. These software maintenance events, however , have a cost – they take a fighter aircraft out of active service for the duration of the maintenance action. While software may not be the root cause of all such software maintenance events, if it were responsible for some, then perhaps models could be developed to predict their frequency. With these models, the software organization could then drive their software development efforts toward a specifiable and measurable operational quality goal and later observe the results of that effort in the field. This article describes a practical and tested approach to using SRE time-based software reliability growth models to relate the release quality of fighter aircraft avionics software to the occurrence of software maintenance events in field operations. Using this approach, the frequency of expected software maintenance events can be estimated before the software is deployed to the field. This provides the software development organization with a key software quality metric, one with a customer oriented view of the operational performance of the avionics software – a mean time to next software maintenance event. Software is nothing more than a set of instructions and data derived from a software design used to control the operation of a computer system. Once software is written, it is tested in stages (e.g., unit, component, system) to identify and remove defects [1]. All contracted capabilities in software are tested and verified prior to delivery. However, it is nearly impossible with large embedded fighter aircraft real-time computer systems to completely and economically test the …
File Format	PDF HTM / HTML
Alternate Webpage(s)	http://www.crosstalkonline.org/storage/issue-archives/2006/200611/200611-Siok.pdf
Alternate Webpage(s)	http://www.stsc.hill.af.mil/crosstalk/2006/11/0611SiokWhittakerTian.pdf
Language	English
Access Restriction	Open
Content Type	Text
Resource Type	Article