An agent-assisted layered multicast architecture for videoconferencing in heterogeneous internetworks

Multicasting is a bandwidth-efficient approach for transmitting real-time audio and video to multipoint receivers over the current Internet. The heterogeneous nature of the Internet, however, can cause a conflict between receivers in multi-party videoconferencing. In a real-time videoconferencing session with many participants, some receivers could suffer congestion when their link bandwidths are very limited or congested with heavy traffic. Considering the different link capacities, a single multicast stream cannot always provide optimum service to all receivers.; Layered multicast has been proposed as an effective solution to cope with the heterogeneity of network bandwidth for multi-party videoconferencing over the Internet. In layered multicast, a source transmits layered video streams through multiple channels. By adjusting the subscription level of the layers, each receiver can dynamically adapt to its local capacity.; In this research, we developed a new architecture to support layered multicasting for one-to-many videoconferencing over the current best-effort and heterogeneous Internet. A new network function, the Agent, which can be located in an outgoing router connecting a homogeneous network to the Internet backbone, is designed to assist a receiver in its decision to join a higher layer and for its coordination with other receivers.; A dynamic Quality of Service (QoS) control algorithm for adjusting the subscription level of layers at the receiver is proposed to reduce packet loss rate by reacting quickly to network congestion. Other important issues, including intra-session information sharing and inter-session bandwidth control, are also addressed. The new architecture is implemented on an existing videoconferencing tool, CafeMocha, which has been developed in the Multimedia Communications and Networking Laboratory at Texas A&M University. The soundness of the proposed concepts is evaluated by experiments using CafeMocha.