The rapid growth of fixed and mobile communicating objects raises the issue of their integration in everyday environment, e.g. in a e-health or home-automation context. The main challenges of these objects are the interoperability and the handling of the massive amount of data that they generate. Our objective is to develop an open, interoperable machine to machine framework maintaining the recently elaborated standards. This framework will allow the realization of services based upon connected devices. To insure the efficient development of that framework, an abstract model of connected objects needs to be specified, along with the elaboration of a formal algebra describing the composition of the connected devices, the handling of the important amount of data they will generate and finally it is of major importance to understand and model their cyber-physical proprieties. In order to complete the development successfully, three research directions will be followed: the specification of a connected device model based on its cyber-physical properties along with a service composition algebra; being able to insure a smart self-adaptation in a changing context allowing the connected devices to reason about themselves, the handling of the events flow, which requires the definition of an event model to make the objects interoperable; and finally the definition of a declarative language allowing the detection of significant events in the massive generated data. The testing of these contribution will be realized through the development and the large scale experimentation of a remote monitoring and diagnostics service based on the mass collection of medical data with non invasive embedded biomedical sensors network.