The first characteristic which makes multimedia data different from classical computer data (text, binary data, …) is their computing unit: multimedia data consist of flows, while texts and binary data are handled as files. Indeed, audio and video data are sequences of images or audio samples which succeed each other with a constant or not rate. Multimedia data is not incompatible with the notion of computer file, and files are still used to save movies or audio documents; but the computing are made information unit by information unit (image by image, for instance); these information units, back to back, form a flow. Note however that texts, graphics and binary data can take advantage of this kind of computing.However, flows are also characterized by the temporal relationships which exist between the different information units. For example, it does not exist any temporal relationship between the characters which form a textual flow. Similarly for fix images which can be considered as bit flows, or for graphics, it does not exist any temporal relationships between the different units of the flow. These are typically discrete flows.At the opposite, for video or audio, images or audio samples must be produced, computed and presented at a regular rate. These are continuous flows (or streams). If the time interval between two consecutive flow units is constant, the flow is said isochronous; nevertheless, a given variability on these time intervals can be tolerated: this variability is called authorized jitter.