Product Reasoning Services: Economic Relevance and Architectural Approaches

Abstract

Numerous business processes require reasoning about goods, services or delivery details. Reasoning is here used as a term for deriving new statements from an existing knowledge base. Examples of such reasoning tasks are the following queries: - Can product 1 be used as a substitute for product 2? - Does a specific product (described by unstructured text data or a vendor-specific number) belong to a given class of a classification system (eCl@ss, UN/SPSC,...)? More sophisticated reasoning could generate statements on transportation requirements or legal aspects. The crucial difference to a mere database query is that the desired answers are not explicitly stored in the system, but are derived from implicit knowledge in the knowledge base. Such reasoning engines will be important components of future eBusiness architectures. This is because for a huge set of transactions, computer systems require answers to such queries, in order to be able to perform the transaction autonomously. Of course, such reasoning engines can be part of the internal logic of appropriate eBusiness applications. There are, however, good reasons (e.g. incentive conflicts, reusability issues, and the need for global availability) to implement them as independent components of eBusiness architectures. For methods that need to be remotely invoked in a distributed environment, Web Services seem to become the architectural component of choice. It is thus reasonable to implement the respective reasoning engines as SOAP-based Web Services. The paper identifies appropriate reasoning services, discusses their internal structure, suggests suitable method calls and their parameters, and explains the implementation as a SOAP-based RPC-style Web Service.