Reducing Efforts in Web Services Refactoring

Abstract

In Service-Oriented Computing, systems provide a service interface described by a computer-readable language called Web Services Description Language (WSDL), but document descriptions often exhibit design problems as systems expand. Moreover, a major problem in this type of applications is its growth; as size and complexity of applications increase, the probability of duplicity of code also increases. This issue could have a negative impact on quality attributes, such as performance, maintainability and evolution, among others, providing developers with some clues to detect refactoring opportunities. Conducting a detection process of these opportunities could be a daunting task; however, this work proposes a methodology to conduct manual refactoring of service descriptions of legacy systems. The methodology allows software developers to collect metrics of time effort and space reduction. The metrics indicate a 75% reduction in services and 63% of the lines of code on average. Additionally, a statistical analysis carried out with the obtained metrics yielded as result that the number of services refactored according to the size of its refactoring group is adjusted to a logarithmic function. We also analyzed how refactoring time increases with the size of services. The results indicate that our methodology could be applied in similar case-studies based on legacy systems.