CFP: Integration of Fountain Codes and Optimal Probabilistic Forwarding in DTNs

There has been much research focusing on the routing problem in delay tolerant networks (DTNs). Much of the work has mainly focused on coding schemes for message distribution, while other work has been done on the probabilistic forwarding. Coding schemes achieve higher delivery rate via redundancy before forwarding, while probabilistic forwarding efficiently limits the abuse of the store and forward scheme, maintaining relatively high performance. Providing a reliable and efficient forwarding scheme proves to be challenging as coding and forwarding schemes should be jointly considered. In our paper, we present an optimal probabilistic forwarding scheme using fountain code, which we name as CFP (Coded Forwarding Protocol), where reliability and efficiency can be achieved at the same time. In CFP, We use fountain codes to encode messages and provide the forwarding rule to decide whether to forward messages to another node. The probabilistic forwarding problem is modeled as an optimal stopping problem, and our forwarding rule also considers the influence of fountain codes. We perform trace-driven simulations and compare CFP with other protocols. Simulation results show that, considering the delivery rate, delay, and number of forwardings, CFP performs better than other implemented protocols - Epidemic, which is the most original protocol, and OPF, which represents the optimal probabilistic forwarding protocol - in our simulation.