To demonstrate a proof of concept of our design work, we implemented SCUBA in both our video conferencing tool, vic [16], and our our media gateway, rtpgw [2]. These implementations follow the algorithms detailed in Section 2.2 and illustrated in Figure 3. In conjunction with our SCUBA-enhanced vic, we deployed and controlled these gateways in the manner described in Section 4.2. In future work we plan to extend our vic and rtpgw implementations for layered transmission and RLM [17] using the signal mapping algorithm detailed in Section 2.3.
Section 3 detailed the limitations of the sampling approach to calculating the average source weight in low-bandwidth sessions with large numbers of receivers. We are currently developing an alternative approach to sampling that employs a spatial network hierarchical to enhance the scalability of the announce/listen receiver interest protocol. Each level of the hierarchy corresponds to a ``report region'' and within each region, a representative is elected. The representative aggregates the reports within that region, summarizes the aggregate information into a single report, and forwards that report up the hierarchy. In this way we greatly enhance the scalability of the protocol since the average weight convergence time is approximately proportional to the number of sources in local regions rather than in the entire session.