Over the past twenty years, service-oriented computing has promoted interoperability between distributed systems
with the help of XML-based languages and protocols (SOAP, WSDL, UDDI). Typical service-oriented architectures (SOAs) rely on centralized solutions for typical service tasks such as discovery and composition. Despite their success in the entreprise world, SOAs have suffered from different problems such as lack of semantic interoperability and scalability issues, that hampered their adoption on a large scale and on the Web in particular, opening new challenges for the research community.
At the same time, the semantic Web has leveraged the publication of data sets that refer to each other over the Web, leading to a distributed dissemination of data (linked data) accessible via RESTful APIs. Nowadays, the convergence of the semantic Web and Web services into a single framework enabling read/write access to Linked Data is an ongoing research challenge for the research community and normalization organizations ( http://www.w3.org/2012/ldp/charter). Most upcoming solutions rely on RDF-based services described and available through Web APIs to access data. These services are called linked services as they are described with and deliver linked data, sometimes through RESTful APIs. The mechanisms and algorithms for discovery and composition typically adopted
The mechanisms and algorithms for discovery and composition typically adopted in SOAs need to be revisited to adapt to the distributed setup of the semantic Web that requires scalable solutions. As well, service descriptions must be adapted to make the semantics of services explicit, fulfilling the needs of SOAs, and as well to integrate seamlessly.