Layered Multicast Rate Control
Research Overview/Summary of Contributions:
Our research has shown that globally optimal rates for layered multicast traffic can be attained in a scalable manner through distributed coordination. Our first paper on this topic [1] has received 65+ citations so far, and has fostered subsequent research on this issue. Our work on this topic extends and nicely generalizes the seminal optimization based rate control work of Kelly et al. and Low et al., to multirate (layered) multicast traffic flows. One of our solution approaches [2] uses subgradient penalty function based non-linear optimization method to develop a primal based distributed optimal rate control algorithm. Our latest and best solution approach for this problem [3] – uses an elegant combination of dynamic programming and Lagrangian relaxation to attain the optimal multicast rates using local message passing, with a convergence speed that is an order of magnitude better than that of earlier solution approaches. Our solution mechanism enables fair and efficient bandwidth sharing by real-time audio and video traffic over the Internet.
The key research papers on this topic are:
Other papers on this topic can be found on my full publication page.