How does selfishness influence the performance of energy-constrained gossip?

In disruption tolerant networks, gossip is an efficient routing scheme, which significantly reduces the message delivery overhead while maintaining a relatively high delivery rate. This mechanism needs network nodes to forward messages and data according to the system-defined gossip probability in a selfless and cooperative way. In this paper, we investigate how node selfishness influences the performance of energy-constrained gossip. By modeling the data transmission process with selfish behaviors in the gossip using a two-dimensional continuous time model of Markov chain, we derive closed-form formulae for average data delivery delay and cost. Extensive numerical results show that gossip is robust enough to selfish behaviors since even when they increase the data delivery delay, there is a gain on the data transmission cost. Moreover, we analyze tradeoffs among the gossip probability, selfishness and energy constraint on the system performance.