Social Groups Based Content Caching in Wireless Networks

Abstract

The unprecedented growth of wireless mobile traffic, mainly due to multimedia traffic over online social platforms has strained the resources in the mobile backhaul network. A promising approach to reduce the backhaul load is to proactively cache content at the network edge, taking into account the overlaid social network. Known caching schemes require complete knowledge of the social graph and mainly focus on one-to-one interactions forgoing the prevalent mode of content sharing among circles of 'friends'. We propose Bingo, a proactive content caching scheme that leverages the presence of interest groups in online social networks. The mobile network operator (MNO) can choose to incrementally deploy Bingo at select network nodes (base stations, packet core, data center) based on user profiles and revenue numbers. We approximate the group memberships of users using the available user-content request logs without any prior knowledge of the overlaid social graph. Bingo can cater to the evolving nature of online social groups and file popularity distribution for making caching decisions. We use synthetically generated group structures and simulate user requests at the base station for empirical evaluation against traditional and recent caching schemes. Bingo achieves up to 30%-34% gain over the best baseline.