Tesi magistrale in Informatica, Dipartimento di Matematica, Scuola di Scienze, UniPD.

Abstract

The Internet of Things is an especially prominent vector of evolution for commercial applications, including smart-home ecosystems. An IoT ecosystem is an ensemble of web-connected devices able to collect, send and act on environmental data. Individual manufacturers employ proprietary data models to endow their devices with ad-hoc properties, functionalities and relationships.

When a 3rd-party ecosystem integration is required in a product, a common language needs to be defned, overarching foreign data models, so that all devices may be interacted with. To address this need, an integrator must use a local or remote middleware, whose deployment afects response times, user functionalities, and maintainability.

The former option is less general but incurs less latency; the latter is more versatile, at the cost of higher latency and more complex data exchange. This thesis surveys and compares state-of-the-art integration strategies, and formulates two original solutions.