Canonical Models for API Interoperability

Canonical Models for API Interoperability

Friday, 11:55am - 12:45pm

As enterprise teams move to adopt REST and modern API management, they are likely to encounter many of the same challenges that hindered, and sometimes derailed, SOA and earlier integration efforts. In particular, interoperability between systems is still a huge challenge for application developers, because each APIs uses its own vocabulary to describe the same essential business domain concepts.


A canonical data model is an ideal solution for this, often recommended, but rarely implemented effectively. Why is it so hard? And why doesn’t REST make these problems go away?


In this session, we’ll break down the challenges, and show how a shared canonical model can make life easier for API consumers, while still allowing the flexibility to expose different services, with different contextual requirements. We’ll explain why the right level of abstraction is so important for canonical models to address the pervasive problem of variability across APIs. We’ll formalize a realization mapping between canonical models and message definitions, and bring this together into a unified client SDK that works across services.

Ted.Epstein's picture
Ted Epstein has over 20 years of experience in commercial and enterprise software development, and has been working on model-oriented solutions for API design and large-scale software integration since 2006. In 2010, Ted left his position at Morgan Stanley to found ModelSolv, a company dedicated to providing innovative technologies and solutions in enterprise API management. Ted has presented at the EclipseCon and Eclipse Banking Days conferences, and is co-organizer of the API-Craft NYC Meetup. @tedepstein