Research Overview

The process of transforming monolithic legacy applications into semantic SOAs requires new research in several areas which constitute the three major research objectives of TAO:

SWS bootstrapping via semi-automatic acquisition of domain ontologies. Semantic descriptions of web services rely on the use of two kinds of ontologies - generic Web Service ontologies and domain-specific ones. While there is much ongoing effort dedicated to generic ontologies (e.g. WSMO), acquisition of domain ontologies is left to the user. Here the objectives of TAO are:

- To identify the major characteristics of the data sources required for learning SWS domain ontologies;choose the right extraction methods.

- Support human corrections of automatically learned domain ontologies as part of the application transitioning process.

Augmentation and integration of legacy content (databases and documents used by applications) relative to domain ontologies to enable ontology-based information access. A new heterogeneous knowledge store is being developed to support scalability and heterogeneity. Efficient support for a combination of structured/semantic queries and keyword search queries is currently partially available only from the leading RDBMS vendors and missing among semantic repositories. In addition, there are no efficient implementations of truly distributed semantic repositories. The state-of-the-art is limited to the distributed ontology meta-data repositories (e.g. Oyster), which only handle Dublin core-like attributes about whole ontologies. This is more than an engineering task, because there are a number of research issues that need to be solved too, e.g., querying distributed repositories.

Transitioning Methodology and Infrastructure. From the integration perspective, not only integration is required, but also new ways to deal with the semantic interoperability among different solutions (new SSOA approaches vs. legacy systems). Research in TAO is tackling these challenges by providing a highly innovative infrastructure for transitioning legacy applications to semantic- and service-based ones via SWS bootstrapping. The main innovative aspect of the transitioning environment is the introduction of automation when supporting the developer in creating SWS definitions, based on the semantic analysis of existing application documentation and legacy content. The TAO infrastructure is based on and compliant with relevant architectures, such as WSMO, WSMX, SWSI, and supports ontology-based legacy data integration.

These three objectives can be thought of as a prism refracting legacy systems into semantic SOAs:

tao transitioning

The project's objectives are also being validated in two case studies, in order to prove the cross-domain applicability of the work. The two case studies play two very different roles - one is to be an open source reference application, forming the basis of dissemination, whereas the other one forms the basis for commercial exploitation. This is aimed to maximise outreach and scientific impact, while at the same time maintaining strong commercial potential.

The first case study undertakes the transitioning of an existing large-scale open-source system, its software documentation, related papers, video tutorials, etc. The resulting ontology- and service-based system acts as a publicly available reference showcase of the TAO technology, demonstrating the advantages offered by semantic SOAs.

The second case study focuses on transitioning the production process of aircraft maintenance documentation. Ontology-based technologies improves the current labour-intensive knowledge production process by adding automation and a collaborative element through SWSs. In addition, TAO enables exploitation of the documentation-related applications by other business units through the composition of SWS.