We describe a system for multipoint videoconferencing that offers extremely low end-to-end delay, low cost and complexity, and high scalability, alongside standard features associated with high-end solutions such as rate matching and per- sonal video layout. The system accommodates heterogeneous receivers and networks based on the Internet Protocol and relies on scalable video coding to provide a coded representation of a source video signal at multiple temporal and spatial resolutions as well as quality levels. These are represented by distinct bitstream components which are created at each end-user encoder. Depending on the specific conferencing environment, some or all of these components are transmitted to a Scalable Video Conferencing Server (SVCS). The SVCS redirects these components to one or more recipients depending on, e.g., the available network con- ditions and user preferences. The scalable aspect of the video coding technique allows the system to adapt to different network conditions, and also accommodates different end-user requirements (e.g., a user may elect to view another user at a high or low spatial resolution). Performance results concerning flexibility, video quality and delay of the system are presented using the Joint Scalable Video Model (JSVM) of the forthcoming SVC (H.264 Annex G) standard, demonstrating that scalable coding outper- forms existing state-of-the-art systems and offers the right platform for building next-generation multipoint videoconferencing systems. We describe a system for multipoint videoconferencing that offers extremely low end-to-end delay, low cost and complexity, and high scalability, alongside standard features associated with high-end solutions such as rate matching and per-sonal video layout. The system accommodates heterogeneous receivers and networks based on the Internet Protocol and relies on scalable video coding to provide a coded representation of a source video signal at multiple temporal and spatial resolutions as well as quality levels. These are represented by distinct bitstream components which are created at each end -user encoder. Depending on the specific conferencing environment, some or all of these components are transmitted to a Scalable Video Conferencing Server (SVCS). The SVCS redirects these components to one or more recipients depending on, eg, the available network con- ditions and user preferences. The scalable aspect of the video coding technique allows the system to adapt to different networ k conditions, and also accommodates different end-user requirements (eg, a user may elect to view another user at a high or low spatial resolution). Performance results concerning flexibility, video quality and delay of the system are presented using the Joint Scalable Video Model (JSVM) of the forthcoming SVC (H.264 Annex G) standard, demonstrating that scalable coding outperforms forms of state-of-the-art systems and offers the right platform for building next-generation multipoint videoconferencing systems.