MDPE LogostLogotudLogorewerseLogo modelplexLogo
MDPE: Model-Driven Performance Engineering


Welcome to the Model-Driven Performance Engineering (MDPE) website. This site contains information about the development of the MDPE Workbench. Model-Driven Performance Engineering is a methodology to extend existing process modeling tools with functionality for performance engineering. It therefore allows detection of performance issues by deriving performance models from process models in an automatic manner, analyzing the generated performance models by external analysis tools and finally returning the performance feedback directly into the original process development environment.


The major goal of Model-Driven Performance Engineering is to integrate performance engineering techniques throughout the lifecycle of software and its development processes with minimal effort. Performance analysis should be possible at early stages of software development processes and should be possible without performance engineering expertise. Existing process models shall be reused and extended to simulate and predict performance bahaviour of software systems. Thus, the costs, risks and efforts of software development can be reduced and flaws in performance requirements can be detected early.


Although performance is one of the most important non-functional requirements of all software systems, performance often plays no - or only an unimportant - role during the software development process. MDPE aims to help the software developers to predict performance behaviour at early stages of the software development process.

Many performance related questions can be answered or predicted using MDPE. E.g.,

  • The execution time of processes and its variation caused by increasing or decreasing workload,
  • The identification of bottlenecks and sensitive parameters to optimize future process execution,
  • The investigation if a redistribution or alteration of resources would improve the performance.


Model-Driven Performance Engineering is based on the idea to automate performance model creation by using model-driven methodologies such as megamodelling and model transformation. The input data of MDPE are existing process models that can be annotated with additonal resource and requirements data. Afterwards, performance models are transformed on the existing, enriched process models and are then used as input models of different performance analysis tools to predict performance behaviour.

The MDPE data flow as a Block Diagram.

The MDPE Workbench supports the combination of different process modeling tools with different analysis tools. Furthermore, the MDPE Workbench can combine several analysis tools to (semi-)automate performance analysis to Analysis Tool Chains.