CO-RTO: Achieving efficient data retransmission in VNDN by correlations implied in names

Abstract

Packet loss is a serious phenomenon in Vehicular Ad-hoc NETworks (VANETs). Although the multi-source and in-network caching characteristics of Named Data Networking (NDN) can help the consumer to retrieve named content, how to provide reliable named data transmission in vehicular environment is still a problem. To achieve the named data retransmission, previous works follow the traditional RTO algorithms based on the latest RTT samples. Since the host-to-host transmission mode is broken by the multi-source and high dynamic Vehicular Named Data Networking (VNDN), those algorithms are no longer applicable. In this paper, we first put forward the concept of correlativity between two chunk transmissions, then quantify it by the logic and physical semantics implied in names, and finally propose CO-RTO, a correlativity-based algorithm which estimates the Retransmission TimeOut (RTO) according to the RTT samples with high correlativity. We have validated our approach in an urban traffic scenario on ndnSIM. The simulation results show that CO-RTO is more efficient than some classic algorithms in named data retransmission.