The emergence of interactive surfaces and technologies able to differentiate users allows the design and development of Identity-Aware (IA) interfaces, a new and richer set of user interfaces (UIs). Such user interfaces are able to adapt their behavior depending on who is interacting. However, existing implementations, mostly as software toolkits, are still ad-hoc and mostly based on existing GUI toolkits which are not designed to support user differentiation. The problem is that the development of IA interfaces is more complex than the development of traditional UIs and still requires extra programming efforts. To address these issues, we present a set of implementation models, named IOWAState models, to specify the behavior as state machines, the architecture and the components of IA interfaces. In addition, based on our IOWAState models and a classification of IA user interfaces, we detail a set of design patterns to implement the behavior of IA user interfaces.