Software Defect Analysis of a Multi-release Telecommunications System

This paper provides a study of several process metrics of an industrial large-scale embedded software system, the Lucent product Lambda-Unite MSS. This product is an evolutionary hardware/software system for the metropolitan and wide-area transmission and switching market. An analysis of defect data is performed, including and comparing all major (i.e. feature) releases till end of 2004. Several defect metrics on file-level are defined and analyzed, as basis for a defect prediction model. Main analysis results include the following. Faults and code size per file show only a weak correlation. Portion of faulty files per release tend to decrease across releases. Size and error-proneness in previous release alone is not a good predictor of a file's faults per release. Customer-found defects are strongly correlated with pre-delivery defects found per subsystem. These results are being compared to a recent similar study of fault distributions; the differences are significant.

[1]  Steel T. Huang,et al.  Statistical modeling applied to managing global 5ESS®-2000 switch software development , 1997, Bell Labs Technical Journal.

[2]  Dewayne E. Perry,et al.  A case study in root cause defect analysis , 2000, Proceedings of the 2000 International Conference on Software Engineering. ICSE 2000 the New Millennium.

[3]  Elaine J. Weyuker,et al.  Where the bugs are , 2004, ISSTA '04.

[4]  Dewayne E. Perry,et al.  Classification and evaluation of defects in a project retrospective , 2002, J. Syst. Softw..

[5]  Mary Shaw,et al.  Empirical evaluation of defect projection models for widely-deployed production software systems , 2004, SIGSOFT '04/FSE-12.

[6]  Elaine J. Weyuker,et al.  A Tool for Mining Defect-Tracking Systems to Predict Fault-Prone Files , 2004, MSR.

[7]  Audris Mockus,et al.  Understanding and predicting effort in software projects , 2003, 25th International Conference on Software Engineering, 2003. Proceedings..

[8]  Jeff Tian.  Quality-evaluation models and measurements , 2004, IEEE Software.

[9]  Norman E. Fenton,et al.  Quantitative Analysis of Faults and Failures in a Complex Software System , 2000, IEEE Trans. Software Eng..

[10]  P. Hartman.  Utility of popular software defect models , 2002, Annual Reliability and Maintainability Symposium. 2002 Proceedings (Cat. No.02CH37318).

[11]  Giovanni Denaro,et al.  An empirical evaluation of fault-proneness models , 2002, ICSE '02.

[12]  Robert E. Park,et al.  Software Size Measurement: A Framework for Counting Source Statements , 1992 .

[13]  Elaine J. Weyuker,et al.  The distribution of faults in a large industrial software system , 2002, ISSTA '02.