Load-Efficient Maximum-Throughput Peer-to-Peer Multicast in the Internet

Data distribution in the existing peer-to-peer multicast protocols is performed from source node by delivering data to multicast nodes over a tree, multi-tree, or mesh path structure covering all the multicast nodes. The existing algorithms try to maximize throughput, minimize latency, or minimize load on network but not to achieve all the three goals at the same time. In this paper, we propose a network optimization problem called LEMT to minimize the load imposed on the network while achieving the maximum multicast throughput. We divide LEMT into a number of partial-LEMT problems and propose a polynomial-time algorithm called PLDD that finds an approximated solution for LEMT. PLDD is able to adapt to heterogeneous multicast groups and needs a node (such as the source) aware of global topology to schedule block transfers between the nodes. Comparison to the existing mesh-based algorithms shows improvement on multicast throughput, load, and latency.