ODC - a 10x for Root Cause Analysis
Procedings RAM 2006 Workshop, Berkeley CA, May 6, 2006
ODC - a 10x for Root Cause Analysis
Ram Chillarege
Chillarege Inc.
April 2006 ram@chillarege.com, +1 (917) 790 9390, www.chillarege.com
Abstract -- Orthogonal Defect Classification (ODC) allows us to do a “10x” on Root Cause Analysis
(RCA). It is a 10x in terms of the time it takes to perform root cause analysis and a 10x in terms of the coverage on the defect stream. These productivity enhancements are achieved by raising the level of abstraction and systematizing the analysis methodology. The impact of this productivity boost is farreaching in its business impact with reported gains that are enormous.
Introduction
Root cause analysis is the staple of productivity improvement methods in software engineering. Its success is legendary but the degree of implementation is often shallow. The difficulties in implementation are tied to cost, coverage, and skills needed to execute an effective root cause analysis. This puts an enormous strain on the manager who wants to use root cause analysis to understand issues and take action, but is unable to do so with speed and simplicity. As a consequence many opportunities which exist for quality improvement remain unexplored and unexploited.
The benefits of root cause analysis are usually quite obvious but the work load it generates is not as obvious.
As a consequence it is not surprising to see institutions that begin such an exercise with gusto only to see the efforts diminish with time. Thus benefits are reaped only initially and the methodology and practice gradually fade from the organization.
ODC changes the game of how root cause analysis is done. It shifts the workload from detailed analysis of individual defects to a rapid categorization process that extracts the semantics from defects without detailed defect root cause analysis. Groups of defects are then analyzed using a measurement model that relates causeand-effect. The
ODC - a 10x for Root Cause Analysis
Ram Chillarege
Chillarege Inc.
April 2006 ram@chillarege.com, +1 (917) 790 9390, www.chillarege.com
Abstract -- Orthogonal Defect Classification (ODC) allows us to do a “10x” on Root Cause Analysis
(RCA). It is a 10x in terms of the time it takes to perform root cause analysis and a 10x in terms of the coverage on the defect stream. These productivity enhancements are achieved by raising the level of abstraction and systematizing the analysis methodology. The impact of this productivity boost is farreaching in its business impact with reported gains that are enormous.
Introduction
Root cause analysis is the staple of productivity improvement methods in software engineering. Its success is legendary but the degree of implementation is often shallow. The difficulties in implementation are tied to cost, coverage, and skills needed to execute an effective root cause analysis. This puts an enormous strain on the manager who wants to use root cause analysis to understand issues and take action, but is unable to do so with speed and simplicity. As a consequence many opportunities which exist for quality improvement remain unexplored and unexploited.
The benefits of root cause analysis are usually quite obvious but the work load it generates is not as obvious.
As a consequence it is not surprising to see institutions that begin such an exercise with gusto only to see the efforts diminish with time. Thus benefits are reaped only initially and the methodology and practice gradually fade from the organization.
ODC changes the game of how root cause analysis is done. It shifts the workload from detailed analysis of individual defects to a rapid categorization process that extracts the semantics from defects without detailed defect root cause analysis. Groups of defects are then analyzed using a measurement model that relates causeand-effect. The
References: and Kothanda Ram Prasad, Proceedings International Conference on Dependable Systems and Networks, IEEE, 2002. IBM Systems Journal, Vol 41, No. 1, 2002. Chillarege, Zbigniew Kalbarczyk, IEEE Transactions on Dependable and Secure Computing, Vol No.2, 2004. Intl. Symposium on Software Reliability Engineering, IEEE, Boca Raton, FL, Nov 4, 1999.