The capacity of epidemic routing in vehicular networks

In this letter, we present an analytic framework to evaluate the capacity of epidemic routing in vehicular networks. Using this analytic model, we derive the upper and lower-bound of end-to-end delay in epidemic routing. Our analytic results reveal that epidemic routing achieves maximum diversity gain over unicast routing when the vehicular network is partially connected, which is in contrast to the expectation that epidemic routing performs better in disconnected networks. We validate the accuracy of our model through simulations.