Sup­port­ing the De­riv­a­tion of Com­pre­hens­ive Spe­cific­a­tions for Black-box Ser­vices

Many services that exist today it cannot be accurately determined if they can be reused in new contexts because of the lack of comprehensive service specifications. In this project, semi-automatic techniques are investigated that provide support to derive comprehensive specification that capture ontological and behavioral service semantics and a service grounding.

Members: Simon Schwichtenberg

Contact: Simon Schwichtenberg

External Funding: was partially supported by the German Research Foundation (DFG) within the Collaborative Research Centre “On-The-Fly Computing” (SFB 901)

Description:

Today, cloud vendors host third party black-box services, whose developers usually provide only textual descriptions or purely syntactical interface specifications. Cloud vendors that give substantial support to other third party developers to integrate hosted services into new software solutions would have a unique selling feature over their competitors. However, to reliably determine if a service is reusable, comprehensive service specifications are needed. Characteristic for comprehensive in contrast to syntactical specifications are the formalization of ontological and behavioral semantics, homogeneity according to a global ontology, and a service grounding that links the abstract service description and its technical realization. Homogeneous, semantical specifications enable to reliably identify reusable services, whereas the service grounding is needed for the technical service integration.

In general, comprehensive specifications are not available and have to be derived. Existing automatized approaches are restricted to certain characteristics of comprehensiveness. This project considers an automatized approach to derive fully-fledged comprehensive specifications for black-box services. Ontological semantics are derived from syntactical interfaces specifications. Behavioral semantics are mined from call logs that cloud vendors create to monitor the hosted services. The specifications are harmonized over a global ontology. The service grounding is established using traceability information. The approach enables third party developers to compose services into complex systems and creates new sales channels for cloud and service providers.