Hybrid Model for Antipattern Analysis in Web Services

Abstract

Performance of service based systems depends on the architectural design of the web service architecture. Presence of Antipatterns in the system lead to maintainability issues and impact the overall quality of service computing architecture. Quality measurement of the service based architecture can be done using metric analysis using well defined thresholds. A variety of static metrics based detection approaches are defined in the literature but detection using dynamic metrics is still in infancy. The main objective of this paper is to propose a heuristic based hybrid model for Antipattern detection using static, interface and runtime execution metrics. A comparison of the result generated by the newly proposed model is done with the existing algorithm. The results generated using the proposed model clearly depicts that performance of the proposed hybrid model is better in terms of precision and recall rate. Along with that the threshold estimation technique used in the paper leads to generation of more precise threshold which ultimately optimizes the generated results.