On selection of forwarding nodes for long opportunistic routes

Opportunistic routing is a promising routing paradigm which increases the network throughput. It forces the sender's neighbors, who successfully overheard the transmitted packet, to participate in the packet forwarding process as intermediate forwarding nodes. As a seminal opportunistic routing protocol, MORE combines network coding idea with opportunistic routing to eliminate the need for strict coordination among active forwarding nodes. In this paper, we show that MORE performance does not scale well with the route length, especially when the route length goes beyond two hops. Also, we found that MORE fails to establish a working opportunistic route in sparse networks. Clearly, the network throughput is directly influenced by both the quantity and quality of forwarding nodes, and their cooperation order. In this paper, we propose a new forwarder selection mechanism which considers the route length, link qualities, the distance from the source, and nodes density. It eliminates the occasional route disconnectivity happening in MORE and improves the quality of the established opportunistic routes. The simulation result indicates that our proposal always outperforms MORE when dealing with long opportunistic routes.