GR 2813 CORE : ISSUE 1

Preface
1 Introduction
2 General Information
2.1 Software Release
2.2 Software Faults
2.3 Software Failures
2.4 Failure Processes
3 Requirements
3.1 Software Reliability Metrics for Ongoing
                Operations
3.2 Software Reliability Metrics for Installation
3.3 Data Requirements
3.3.1 Failures
3.3.2 Faults
3.3.3 Software Modules
3.3.4 Software Releases
3.4 Software Reliability Prediction Models
Appendix Example Model
A.1 Model Objective
A.2 Data for Model Predictions
A.3 A Poisson Regression Model
A.3.1 Estimation of Parameters
A.3.2 Confidence Intervals for Parameters
References
Bibliography
Glossary
List of Figures
Figure 2-1 Source Code for Different Releases
Figure 2-2 Life Cycle of Software Release
Figure 2-3 Example of Software Release History
Figure 3-1 Data for Software Reliability Prediction
List of Tables
Table 2-1 Application of Software Reliability Prediction
                in the Software Development Cycle
Table 2-2 Data Describing a Single Software Failure for
                Software Reliability Prediction
Table 3-1 Operation Failure Rates for System S
Table 3-2 Installation Failure Rates for System S
Table A-1 Common Functional Forms for Failure
                Intensity l
Table A-2 Grouped Failure Data by Release
Table A-3 Estimated Failure Intensity Parameters for
                Logarithmic Poisson Model Based on
                Maximum Likelihood Estimates

Prediction of reliability is an important element in selecting equipment for use by the Bellcore Client Companies. In the past, such a statement concerning reliability has implied hardware reliability. However, software's role in the operation of telecommunications equipment has become increasingly important. Therefore, this document proposes generic requirements for predicting software reliability. Provides attributes that a software reliability prediction model is required to possess. Suggests a model framework that complies with these attributes and can therefore be used for prediction. To promote the continued development of improved models and to allow current models to be continually validated, this document also give criteria for data collection for supporting software reliability prediction.