While mobile services incorporate and apply the fundamental principles of Service Oriented Architecture (SOA), they present a number of specifics that push certain challenges related to service oriented systems to their extreme and additionally introduce new unique research challenges. Such specifics relate to: (i) dynamism – open mobile environments are much more volatile than typical service environments, with services emerging and disappearing in arbitrary ways without prior notification; (ii) heterogeneity – a direct consequence of ad hoc mobile environments is that no safe assumption can be made about the technological and business features of the services encountered; (iii) awareness – in most mobile service applications, the business capabilities of services are not the only ones that matter, the multi-faceted context of services is equally important; and (iv) the equation among QoS expectations on services, scalability, and required resources is hard to solve, due to the resource constraints that are typical to mobile environments. Dealing with the identified specifics gets even more complex if we consider both traditional computing services and services attached to the physical world by means of sensors and actuators, i.e., Things. In this habilitation thesis, I provide an overview and discussion of my main research results in these last years. In the context discussed above, I have focused on two principal aspects of service oriented computing in mobile environments: Interoperability and Composition. In my research work reported herein, Interoperability refers to the middleware layer, while Composition refers to the application layer. More precisely, service Composition relies on a common underlying middleware that supports Interoperability independently of the development and deployment internals of services (hardware platforms, operating systems, programming languages); however, when heterogeneous middleware is employed, Interoperability needs to be worked out also at this layer. Particularly in the open mobile environment reaching out to the whole Internet, these two aspects are closely interrelated. This calls for extensions to the classic SOA architectural style: services can be very diverse networked entities (e.g., Things, sensor-actuator networks, service feeds, data streams). Hence, stating this briefly, my research contributions have aimed at extending the SOA style with Interoperability and Composition enablers for mobile computing.