I attended a super presentation this morning, that is a presentation on Semantics Utilized for Process management within and between Enterprises (SUPER). SUPER’s goal is to use ontological frameworks to make web services autodiscoverable and autoconfigurable.
While web services have enabled interactions between systems and companies more flexibly than ever before, they are still too hard to work with. Today’s web services were often created any naming standards were readily available. All too often, the similar items will have different names in very similar service offerings (poor semantic interoperability). When elements in two services share the same name, they may mean different things to the two services because they mean different things to the two companies (poor ontological interoperability).
Even when the fields have the same definition, they may be laden with external information that is not part of the schema. Certain aparantyl simple terms such as “shipping date” and “receiving date” may be overlaid with legal contracts that are not apparent in the web service description.
The SUPER project attempts to do the semantic assembly of business projects from existing services. By comparing ontologies, SUPER can rank available services as to how well they conform to the expectations of the business analyst. The goal of SUPER automated integration. Today, it can prepare a semi-automated response that leverages annotation, service comparison, and even composing hybrid services to deal with heterogeneity of service offerings.
Beware, though, as SUPER relies on a whole alphabet soup of standards.
SUPER uses Web Service Modeling Ontology (WSMO) to characterize systems, WSMO looks for strict WS compliance and strict decoupling of services. WSMO relies on ontology based classification of services to define core mediation services and execution semantics. WSMO is about service description, not implementation, defining a path to Semantic Business Process Management (SBPM).
SUPER augments SBPM by defining a semantic execution environment (SEE) based upon an open reference ontology. SEE is an architecture that defines what it is possible to ask for, but that does not specify how to do it, or at what granularity. SEE mediates between ontological representations rather than mediating between different behaviors
If you want to play around with this, there is a lot of open source material you can pull down, including:
- Apache ODE, a BPEL aware web server based on Apache
- WSMO Studio, an extended Eclipse
- WSMX on SourceForge
- SUPER, www.ip-super.org
These guys are bright and driven, and their work is well worth checking out.