PerOpteryx/GuiltBasedAntipatternHandling
This page describes details on the case study for the SoSym 2011 submission.
BRS Model
The BRS model described in the paper can be downloaded here: Datei:PerOpteryx GuiltBasedAntipatternHandling BRS model.zip. You can inspect and analyse the model using the Palladio Component Model tool PCM 3.2, which is freely available.
Guilt-based Process
The calculation of the guilt-based process can be traced in the following Excel file, packaged in a zip file: Datei:PerOpteryx GuiltBasedAntipatternHandling Calculation.zip
Results
The first analysis is to apply the antipatterns-based process without the ranking step, i.e. all the detected antipatterns are solved. The figure below reports our experimentation by applying the antipatterns-based process without the ranking step, and across multiple iterations of the process: the target performance index is the response time of the system and it is plotted on the y-axis, while the iterations of the antipatterns-based process are listed on the x-axis. Single points represent the response times observed after the separate solution of each performance antipattern.
A second analysis has been executed by introducing the ranking process with the confidence criterion, i.e. by deciding that all antipatterns whose guilt degree is lower than a confidence value of 0.5 are discarded. The results are shown below.
The raw results are contained in the following Excel file, packaged as a zip file: Datei:PerOpteryx GuiltBasedAntipatternHandling AllCandidates.zip