Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Reliability Prediction Models for Product Lines. A Brief Overview

: Freimut, B.

urn:nbn:de:0011-n-28582 (105 KByte PDF)
MD5 Fingerprint: c87bdce6bb855a983fe4e24f7a3057b0
Created on: 12.12.2000

Kaiserslautern, 2000, VIII, 14 pp. : Ill., Lit.
IESE-Report, 046.00/E
Reportnr.: 046.00/E
Report, Electronic Publication
Fraunhofer IESE ()

When planning product lines, many criteria have to be taken into account in order to decide, for example, whether a specific feature should be developed for reuse or specifically for a single system. One criterion is the reliability of the entire product line. In order to take this criterion explicitly into account, reliability prediction models are required. These models aim at predicting the reliability of a feature based on the starting situation (e.g., whether the feature has to be implemented from scratch, whether existing code can be adapted or reengineered) and the development approach (e.g., develop for reuse or develop system specific).
This report works towards the definition of a reliability prediction model. Specifically, a brief survey of existing reliability prediction models is given. However, none of these approaches is readily implementable in the context of planning a product line. Therefore, as a first step a measurement program is proposed, in which the reliability of features is characterized. For this purpose, several measures resp. models for reliability are presented together with a list of factors that were reported in the literature to have an impact on the reliability of a software system.