Improving entity linking with ontology alignment for semantic microservices composition

Abstract

Purpose This paper aims to propose a method based on Linked Data and Semantic Web principles for composing microservices through data integration. Two frameworks that provide support for the proposed composition method are also described in this paper: Linkedator, which is responsible for connecting entities managed by microservices, and Alignator, which aligns semantic concepts defined by heterogeneous ontologies. Design/methodology/approach The proposed method is based on entity linking principles and uses individual matching techniques considering a formal notion of identity. This method imposes two major constraints that must be taken into account by its implementation: architectural constraints and resource design constraints. Findings Experiments were performed in a real-world scenario, using public government data. The obtained results show the effectiveness of the proposed method and that, it leverages the independence of development and composability of microservices. Thereby, the data provided by microservices that adopt heterogeneous ontologies can now be linked together. Research limitations/implications This work only considers microservices designed as data providers. Microservices designed to execute functionalities in a given application domain are out of the scope of this work. Originality/value The proposed composition method exploits the potential data intersection observed in resource-oriented microservice descriptions, providing a navigable view of data provided by a set of interrelated microservices. Furthermore, this study explores the applicability of ontology alignments for composing microservices.

Open Taiwan Government data recommendation platform using DBpedia and Semantic Web based on cloud computing

“…The LOD cloud (Abele and McCrae, 2017; Assaf et al , 2015) is used to combine distributed and various open data sets into an innovative application, and that it has emerged as one of the largest collections of interlinked open data sets on the internet. In Salvadori et al (2017), the authors propose an approach based on linked data and Semantic Web technologies for composing microservices through data integration.…”

OntoGenesis: an architecture for automatic semantic enhancement of data services

“…Extensional-based techniques can be adapted for service enrichment (Salvadori et al, 2017) and, similarly to OntoGenesis, exploit instance data to find ontology alignments. They rely on individuals obtained from data services and from external sources to check if they have some intersecting properties, i.e.…”

“…Thus, OntoGenesis constructs ontologies with a flat class hierarchy (without subclasses) and the only relations between classes occur through domains and ranges of properties. Although extensional-based techniques can be adapted for service enrichment (Salvadori et al, 2017), such techniques present poor performance (as shown in section 5.5) when data obtained through the service interface is unrelated to data from the external sources and when individuals are not the same. In light of such evidence, we argue that extensional ontology matching algorithms present distinct characteristics from the problem we tackle, specifically property matches.…”

“…Additionally, such services can also employ Semantic Web technologies (Berners-Lee et al, 2001) in order to provide data in a sophisticated machine-readable format and, therefore, be easily reused by third parties and integrated to complex Web applications. Several researchers have been discussing the benefits of employing Semantic Web services; for instance, enhance data interoperability (Salvadori et al, 2017) and assist in automating tasks such as service discovery, selection, composition, etc. (McIlraith et al, 2001;Sycara et al, 2003).…”

“…In contrast, few proposals address the enrichment of data provided by services. For instance, the method proposed by Salvadori et al (2017) aims to enrich representations of data-based microservices with owl:sameAs and rdfs:seeAlso links. A framework aimed to identify alignments between heterogeneous ontologies is also proposed.…”