Watch the detailed propagation of a single multicast message through an eight-node fragment of the network.
Notice how our intelligent MAC control - based on a Response Schedule in the packet header - minimizes channel contention and maximizes bandwidth utilization.
By scheduling the downstream channel access sequence, each node can avoid propagation collisions
Does not incur bandwidth costs of fat contention windows and random backoff periods.
Simple annealing mechanisms shake out "colliding-cousin" nodes
Annealing is automatically promoted up the hierarchy as needed.
All transmissions are verified by handshake (ACK)
Hub to hub propagation saves time by using the (overheard) forwarded message as a free ACK.
This approach is a Big Fun Development innovation, designed especially for the SWARM requirements. It follows certain ideas of Fullmer and Garcia-Luna-Aceves (FAMA), though it tries to reach these through radically different means.
This addresses the problems of contention during the fan-out of a multicast message. There are other contention problems to be addressed as we move forward.