Integrating Linked Data and Services with Linked Data Services

Abstract

Abstract.A sizable amount of data on the Web is currently available via Web APIs that expose data in formats such as JSON or XML. Combining data from different APIs and data sources requires glue code which is typically not shared and hence not reused. We propose Linked Data Services (LIDS), a general, formalised approach for integrating dataproviding services with Linked Data, a popular mechanism for data publishing which facilitates data integration and allows for decentralised publishing. We present conventions for service access interfaces that conform to Linked Data principles, and an abstract lightweight service description formalism. We develop algorithms that use LIDS descriptions to automatically create links between services and existing data sets. To evaluate our approach, we realise LIDS wrappers and LIDS descriptions for existing services and measure performance and effectiveness of an automatic interlinking algorithm over multiple billions of triples.

grlc Makes GitHub Taste Like Linked Data APIs

“…In [13], authors propose to expose REST APIs as Linked Data. These approaches suggest the use of Linked Data technology on top of Web services.…”

Searching Linked Data and Services with a Single Query

“…These services are published in public repositories and are either stored in RDF stores behind SPARQL points or simply as RDF dumps. Many tools already provide that feature (iServe [13] LIDS [18]). The Inputs/Outputs are important parts of the descriptions that should reference to existing ontologies.…”

Rapidly Integrating Services into the Linked Data Cloud

“…The second approach creates services, called Linked Open Services (LOS) [9,5], that consume and produce Linked Data. Finally, Linked Data Services (LIDS) [12,13] integrates data services with Linked Data by assigning a URI to each service invocation. [1] to map service inputs and outputs to the concepts in ontologies.…”

“…A second approach is to wrap the APIs to enable them to communicate at the semantic level so that they can consume and produce linked data [9,5]. A third approach is to create a uniquely identifiable resource for each instance of an API invocation and then link that resource to other data sets [13,12]. Using this approach it is possible to invoke the API by dereferencing the corresponding resource URI.…”