Nowadays software systems have extremely large design space. Dozens to hundreds of parameters comprise the configuration of typical software systems. Palladio configuration optimization is suitable to derive optimal configuration with respect to performance and reliability. Configuration parameters can be systematically evaluated (even combinations of parameters) with respect to their impact of the Quality of Service (QoS) end users can expect from a system which is configured as described in Palladio models.
Opposed to tweaking the configuration of running software systems, one can simply change the model – without re-deploying / re-compiling / re-starting the software system. Hence, Palladio can help saving manual trial-and-error cycles for configuration optimizations.
The essential parameters: Which are the essential and most sensitive parameters to touch when optimizing my software configuration (a so-called sensitivity analysis)?
Parameter values: What are the most suitable parameter ranges? Which parameter values should I choose for the current setup in my specific scenario?
Parameter combination: How do which parameters relate to each other? Which parameter combination does not terminate my optimization effort?