Cost Estimation for Model-Driven Interoperability: A Canonical Data Modeling Approach

Canonical data modeling (CDM) is a method for reducing the total cost of integrating information systems. CDM enhances model-driven development (MDD) by making explicit situations in which data components are being reused (i.e., shared across systems). In this paper, we present cost models for data modeling and data integration, with and without CDM. We then demonstrate how to use those cost models (during project planning) to determine if CDM is a good fit for an enterprise. Next, we provide the first empirical estimates of the costs associated with data component creation, reuse, and integration in an MDD environment. We conclude with technical lessons learned performing CDM for our government clients.