Dynamic replication control strategy for Opportunistic Networks

Abstract

Epidemic routing is proposed as one of the routing protocols for Opportunistic Networks. These kind of networks behave as sparse and/or highly mobile networks in which there may not be a reliable path from source to destination. We study the trade-off between delivery delay/ratio and resource consumption in an Opportunistic Network in which a message has to be spread to each encountered node by epidemic relaying. In addition to the destination, there are several other nodes in the network that can cooperate in relaying the message. We first assume that, at every instant, all the nodes can predict the number of relays storing the message and the number of new message replicators that have received the message. We formulate the problem as a controlled finite discrete Markov chain and derive the optimal closed-loop control (replication policy). However, in practice, the intermittent connectivity in the network implies that the nodes may not have the required perfect knowledge of the system state. To address this greedy issue, we obtain an ordinary differential equation (ODE) approximation for the optimally controlled Markov chain. Finally, we evaluate the performance of the replication control policy over finite networks. Numerical results show that this dynamic replication policy performs close to the optimal closed-loop policy.