Novel activity-based metrics for efficient forwarding over online and detected social networks

摘要

Pocket Switched Networks (PSNs) are challenged wireless networks of handheld mobile devices that exploit user mobility to distribute data in an opportunistic way. In PSNs, a fundamental issue is the choice of a suitable relay node for message forwarding. In such environments, disconnections are very frequent and an end-to-end path from source to destination usually does not exist. Consequently, routing protocols are based on a hop-by-hop communication paradigm. Most of these protocols exploit social behaviour characteristics to decide whether an encountered node is a good relay. Social network information is commonly extracted from encounters detected between mobile devices. However, Internet added online social interaction techniques, which are not based on physical meetings and that reflect users' online behaviour. In this paper we present novel routing metrics for social-based routing in PSNs that exploit both offline and online users' social network information. By proposing a model of dynamic online social network that uses information extracted from offline and online user behaviour, we show that routing centrality metrics, combining node centrality extracted from dynamic online social network and centrality extracted from detected social network, improve delivery ratio and in some scenarios even reduce end-to-end delay.