Classification and Analysis of Static Metric Based Antipattern Detection in Service Computing

Abstract

Background: Design Pattern is regarded as an essential component for enhancement of system design which can further improve the reusability and maintainability whereas antipattern degrades the quality of the program. Antipatterns are sub-optimal implementation choices which initially appears to be a good concept but later on leads to low code maintainability and higher maintenance costs. Objective: The identification of antipatterns which lead to performance degradation plays an important role in the reduction of expensive work and cost involved in maintenance, redesign, reimplementation, and redeployment. The main motivation is to refactor the source code in order to reduce maintenance efforts. Antipatterns impact reliability, testability and maintainability, but they still lack clear identification because of different interpretations and definition of each antipattern. There is a need for a benchmark to analyze the result generated by antipatterns. Methods: Static and dynamic analysis individually do not suffice for antipattern .A hybrid approach is proposed by combining rule based static analysis with dynamic run time analysis. Before applying the hybrid approach a simple manual validation was done to exclude the type of input which had the least probability of containing antipattern. The approach aims at optimizing the results by inclusion of response time metric measure which can be evaluated at runtime execution of the web service. Results and Conclusion: : The paper focusses on detection of antipatterns from web services based on code level and interface level static metrics .Only a brief overview of dynamic approach for detection is proposed. The future scope of this paper will focus on detection of antipattern based on more number of dynamic metrics and also a comparative analysis of the results generated from static, dynamic and hybrid approach.