Semantic Integration of E-Business Schemas via a Canonical Data Model: Assessing the Effort Reduction for B2B Message Exchange

摘要

Despite of research done for decades, B2B integration is still a big cost driver in IT. Standardization efforts could reduce the burden of schema mapping, but the effort for customizing the huge and under specified guidelines increased. Due to the myriad of different needs, different standards coexist, which hinders integration. Instead of pushing companies to adopt huge standards, this paper proposes an iteratively improving schema and mapping derivation system. This approach establishes common business semantics and allows for flexibility, but streamlines companies' integration efforts based on an unexposed, evolving canonical data model. The need for explicit standardization is reduced to a minimum. Our comprehensive evaluation based on real schemas shows a potential to reduce guide creation effort by 50% and mapping effort from 6% to 100%.