A-VIEW: Adaptive bandwidth for telepresence and large user sets in live distance education

For distance education, several obstacles in live video conferencing are preventing widespread adoption. These include network bottlenecks, lack of telepresence, proprietary hardware usage, and inability to scale up for large user sets. In this paper, we present A-VIEW, a novel software system that has an architecture with an extensible multi-display video conferencing and telepresence solution customized for distance education, which can be effectively used for real-time teaching and training. We show sample deployments of A-VIEW with variable number of displays. We contend that a hybrid set of adaptive bandwidth techniques and domain-specific heuristics can be effectively used to accommodate a large number of users. For example, for each client, A-VIEW can operate with a constant bandwidth (upload and download) of around 256 kbps. A-VIEW can also be used at high bandwidths to provide a high-resolution telepresence solution. Unlike many other applications, A-VIEW does not use any proprietary hardware. Having a software only system makes A-VIEW extremely versatile for widespread deployment in teaching and training applications.