Cooperative caching policies in multi-hop wireless networks

Mobile networks have been experiencing an impressive growth in the data traffic in recent years. Video data represented more than half of global mobile data traffic beginning in 2012, indicating that it is having an immediate impact on traffic even today, not just in the future. This enormous traffic growth urges network operators to continuously invest money in enhancing existing wireless infrastructure, whereas users have to spend more money for contents or cloud based services than for the communication access. This dichotomy motivates new network paradigms in which the users’ wireless terminals cooperate to decrease the load in the wireless network infrastructure: the Quality of Experience of the user is improved, even if the network infrastructure has not been upgraded. The cooperation among the wireless terminals motivates towards the novel content-centric networking (CCN) paradigm in which a content-based addressing allows to identify the requested content independently from its location. The content-reuse, content-localization and the large storage capacity in modern mobile devices enable the adoption of cooperative caching schemes, in which each mobile device acts as a caching node. Our contributions lies in proposing a novel caching insertion policy and performing a thorough comparison between different local and distributed caching policies. Our proposed policy is an “interest-based insertion policy” in which user’s interest for a particular content is considered as the main metric of its insertion into the cache. Results obtained suggest that our proposed policy outperforms other stateof- the-art insertion policies. Furthermore, the performance of local caching and distributed caching policies was evaluated under different synthetic request models as well as with real content request traces obtained from an Italian ISP. Comparison results for both fix wireless and mobility traces show that local caching performs better than the more complex distributed policies.