SOA Antipatterns: An Approach for their Specification and Detection. Palma, F., Nayrolles, M., Moha, N., Gu�h�neuc, Y., Baudry, B., & J�z�quel, J. International Journal of Cooperative Information Systems (IJCIS), World Scientific Publishing, March, 2014. 31 pages.
SOA Antipatterns: An Approach for their Specification and Detection [pdf]Paper  abstract   bibtex   
\def§BSSBSs\def§OASOA\def§ODASODA Like any other large and complex software systems, Service Based Systems (§BS) must evolve to fit new user requirements and execution contexts. The changes resulting from the evolution of §BS may degrade their design and quality of service (QoS) and may often cause the appearance of common poor solutions in their architecture, called \emphantipatterns, in opposition to \emphdesign patterns, which are good solutions to recurring problems. Antipatterns resulting from these changes may hinder the future maintenance and evolution of §BS. The detection of antipatterns is thus crucial to assess the design and QoS of §BS and facilitate their maintenance and evolution. However, methods and techniques for the detection of antipatterns in §BS are still in their infancy despite their importance. In this paper, we introduce a novel and innovative approach supported by a framework for specifying and detecting antipatterns in §BS. Using our approach, we specify ten well-known and common antipatterns, including Multi Service and Tiny Service, and automatically generate their detection algorithms. We apply and validate the detection algorithms in terms of precision and recall on two systems developed independently, (1) Home-Automation, an SBS with 13 services, and (2) FraSCAti, an open-source implementation of the Service Component Architecture (SCA) standard with more than 100 services. This validation demonstrates that our approach enables the specification and detection of SOA antipatterns with an average precision of 90% and recall of 97.5%.

Downloads: 0