Semantics are used to mark up a wide variety of data-centric Web resources but are not used to annotate online functionality in significant numbers. That is despite considerable research dedicated to Semantic Web Services (SWS). This has led to the emergence of a new Linked Services approach with simplified and less costly to produce service models, which targets a wider audience and allows even non-SWS developers to annotate services. However, such models merely aim at enabling semantic search by humans or automated service clustering rather than automation of service tasks such as discovery or orchestration. Thus, more expressive solutions are still required to achieve automated discovery and orchestration of services. In this paper, we describe our investigation into combining the strengths of two distinct approaches to modeling semantic Web services “lightweight” Linked Services and “heavyweight” SWS automation - into a coherent SWS framework. In our vision, such integration is achieved by means of model cross-referencing and model transformation and augmentation.
The past decade has seen a range of research efforts in the area of
Semantic Web Services (SWS), mainly aiming at the automation
of Web service–related tasks such as discovery, orchestration or
mediation via broker-based approaches. Building on formal
service semantics, several frameworks, such as SAWSDL [
In the Web context semantics are used to mark up a wide variety of data-centric resources but are not used to annotate online functionality in any form in significant numbers. The reasons for this are two-fold. Firstly, SWS research has for the most part targeted WSDL/SOAP-based Web services, which are not prevalent on the Web. Secondly, due to the inherent complexity required to fully capture computational functionality, creating SWS descriptions has represented an important knowledge acquisition bottleneck and has required the use of rich knowledge representation languages and complex reasoners. There exists an inherent conflict between the need to capture comprehensive and meaningful service semantics – to allow reasoning-based automation of any sort – and the requirement to keep the costs for providing services descriptions low in order to simplify the modeling process and to ensure that efficient and scalable solutions can be implemented. Hence, despite considerable amount of research dedicated to the SWS visio, so far there has been little take up of SWS technology within non-academic environments.
The prevalent lack of impact of SWS technology is particularly
concerning since Web services as such are in widespread use
throughout the Web nowadays, where applications use distributed
HTTP requests via rather lightweight interface technologies such
as RESTful services, HTTP GET-style request or XML-feeds.
Hence, the SWS challenges are of increasingly crucial importance
for today’s highly distributed Web applications. These issues led
to the emergence of more simplified SWS approaches to which we
shall refer here as “lightweight”, such as WSMO-Lite [
In order to tackle the introduced challenges, we aim at combining the two distinct SWS representation approaches (R1) lightweight Linked Services, and (R2) heavyweight SWS descriptions.
While both approaches partially share common schema entities, e.g. both cover aspects such as interfaces and non-functional properties of services, they differ significantly in certain other aspects, for instance, the way the service models are being produced, the nature of the actual produced models or the kind of reasoning facilitated by each approach. For instance, while (R1) is being produced collaboratively as a joint effort by a potentially large group of service providers and consumers, it allows to consider a range of perspectives on one particular service and to gather annotations and RDF-model references to a wide range of existing RDF vocabularies. Hence, they can be described as multifaceted, deliberately incomplete and incoherent. In contrast, the models usually subsumed under (R2), e.g. WSMO-based service specifications, reflect the perspective of one particular SWS provider and describe a service following a meta-model which aims at exhaustive modeling of a service in terms of its core identifying aspects, such as its capabilities or behavioral characteristics. Here, one strives for a much greater level of expressivity and detail and particularly takes into account execution-related aspects. Therefore, such descriptions could best be described as comprehensive, potentially complex and coherent. Developers Applications annotate & reuse services (a) Light-weight Service Annotations (1) referencing (2) transformation request goals (b) Semantic Web Service Execution Environment Web Service
Web Service
Web Service
Web Service
Web Service Fig. 1. From lightweight service annotations to heavyweight
Depending on the quality of the produced service models, the representational approach (R2) facilitates reasoning that allows for automation of certain service-related tasks such as discovery or orchestration but are costly to produce. In contrast, models as in (R1) are less intricate, but also allow only limited reasoning, such as clustering of services or structured searches by humans. While these approaches currently co-exist without a well-defined relationship, we propose two different bi-directional correlations, which are under investigation: (C1) service model cross-referencing, (C2) service model transformation and augmentation. Under (1), we subsume all kinds of references between models across (a) and (b) as depicted in Fig 1. For instance, a lightweight service annotation could point to a heavyweight SWS description that models the same service more exhaustively or vice versa. That would allow semantics to be exploited in (a) as well as (b) for reasoning of different sorts, for instance, to perform some clustering based on (a) to reduce the amount of potentially interesting services for a given query in (b). In addition, (2) considers the transformation between models across (a) and (b), either manually or (semi-)automatically.
We have described a two-stage approach to semantic service representation. By integrating collaborative and user-driven Webscale service annotations with comprehensive SWS specifications, application developers benefit from both low cost for providing annotation and a high level of automation. In that, while taking advantage of service models produced by a large non-expert audience, both structured search for service instances by humans as well as automation of service tasks is supported. In our vision, integration between lightweight service annotations and comprehensive SWS specifications is achieved by different means of (a) model cross-referencing and (b) model transformation and augmentation. While the current solution provides an overall framework for integrated service models which support different levels of automation, future work needs to address the investigation of automated model transformation mechanisms in order to support the seemless integration of instances across distinct service models schemas. Besides, future work needs to investigate the effort required to populate the introduced knowledge bases and the level of automation which is supported.