Modeling the Internet of Things

Abstract

As we extend the reach of the Internet through sensing and automation, networked interactions become more critical and safety and security require improved means for designing and testing system components, platforms and services. A tremendous amount of messy, heterogeneous data passes through such networks; integrating and applying that data so that it can be used for seamless interoperability and higher level reasoning requires conversion into some higher abstraction. This necessitates a new foundation for information modeling and model management, broadly construed. Such a foundation must support model construction in a wide range of formalism as well as capabilities for integrating models across different formalisms. We must also allow for easy model evolution, simplifying both iterative design and long-term maintenance. In this paper, we argue that category theory, a branch of abstract mathematics, provides a firm conceptual foundation for information modeling which already meets most of these criteria. We close with a discussion of some challenges for exploiting category theory in applied contexts.