Ranking Strategies for Quality-Aware Service Selection

Service oriented computing enables the development of applications as compositions of basic entities called services. These services offer business functions, which are used as primary criteria in the service selection algorithms. In distributed scenarios, a large number of services can offer similar functionalities, motivating the embracement of quality attributes as fundamental selection elements. Different strategies can be used to classify services based on these attributes and it is quite difficult to elect one that seems adequate in every scenario. In this context, we present DSOA platform, which is an environment that supports dynamic service compositions based on quality attributes. Particularly, our focus in this paper is in its Service Selection component which supports the definition of new selection strategies at runtime. In order to validate our solution, we perform an experimental evaluation that shows the utilization of different service selection strategies and their performance impacts and effectiveness in selecting services.