Surveillance systems include a large set of techniques for both low level and high level tasks. In particular, in the last decade the research community has witnessed a high proliferation of techniques that span from object detection and tracking to object recognition and event understanding. Although some techniques have been proved to be very effective, those tasks cannot be considered solved. Even less, we can consider concluded the research in the field of the analysis of the activities (event analysis). It is this topic together with the problem of the information sharing among different sensors that represents the core of this work. Here, a system architecture for a video surveillance system with distributed intelligence over multiple processing units and with distributed communication over multiple heterogeneous channels (wireless, satellite, local IP networks, etc.) is proposed. A new real-time technique for changing the video transmission parameters (e.g., frame rate, spatial/colour resolution, etc.) according to the available bandwidth (which depends on the number of the detected alarm situations, on the required video quality, etc.) will be presented.