Contact Prediction based Routing for Opportunistic Networks

Abstract

As nodes experience intermittent connectivity, Opportunistic Mobile Social Networks (OMSN) forward messages via opportunistic contacts. Observed that considering social and contact expectation together would improve an accuracy of prediction, we propose a novel approach called Contact Prediction based Routing (CPR), a quota copy-based routing strategy, it utilizes contact and inter-contact duration to assess the delivery predictability. First, nodes are divided into several communities according to their social similarities. While contact occurs, nodes exchange multiple copies of a message based on their predicted community encounters. This distribution will happen until the remaining copies of the message become one. Then, unlike the traditional routing where each message is stamped with a single forwarder address, the proposed paper augments each message with multiple probable forwarders. The potential forwarders are prioritized based on the remaining-inter contact time between two nodes. Consequently, one of the most possible forwarders decides dynamically to act as a next relay while suppressing others using an acknowledgement scheme. Thus, a heap of retransmission is diminished. The proposed system is implemented in Opportunistic Network Environment (ONE) simulator and performance is analyzed in terms of delivery rate, delay, overhead. The simulation results show that the proposed system diminishes overhead by 85% and the composite metric is increased by 38% as compared to the other routing schemes.