Empirical Analysis of the Performancerelated Risk.pdfVIP

Empirical Analysis of the Performance-related Risk Andreas Schmietendorf*#, Reiner Dumke* * Otto-von-Guericke-Universität Magdeburg, Fakultät Informatik, Institut für verteilte Systeme, Postfach 41 20, D -39016 Magdeburg, E-Mail: {schmiete, dumke}@ivs.cs.uni-magdeburg.de # T-Nova Deutsche Telekom Innovationsgesellschaft mbH, Entwicklungszentrum Be rlin, Wittestraße 30N, D- 13476 Berlin, E-Mail: A.Schmietendorf@telekom.de models and methods of software engineering currently Abstract in use to more detailed analysis in terms of their repre- sentation of SPE-relevant tasks. The current methods of effort estimation frequently take indirect account of the tasks of software perform- The following figure shows typical times when per- ance engineering (SPE), and provide widely differing formance analysis methods are applied. conclusions. In order to create transparency and ac- ceptance for this task that has been growing in impor- Phase 1 Phase 2 Phase 3-5 Phase 6 Phase 7 Real tance for years in the context of the life cycle of IT SIP Pre-project Development Integration Deployment Operation systems, an approach (PRM Performance Risk Model) Performance modelling (business process view) is adopted that derives resource requirements from a Perf