Diversity-improved caching of popular transient contents in Vehicular Named Data Networking

In-network caching, natively enabled by Named Data Networking (NDN), is effective to speed up content retrieval in vehicular environments, where a large variety of contents are typically transient, i.e., they expire after a certain amount of time, and may exhibit different popularity profiles. Lifetime and popularity play a crucial role in the content caching decision: intuitively, caching a popular content with long lifetime can be more useful than caching an unpopular one that is ready to expire and then to be dropped from the content store. At the same time, making nearby nodes caching different contents and, therefore, improving the caching diversity, can be crucial to get better delivery performance over the broadcast wireless medium. In this paper, we devise a novel distributed caching strategy where vehicles autonomously decide which content is to be locally cached according to the content residual lifetime, its popularity and the perceived availability of the same content in the neighborhood. The target is to cache with higher probability more popular contents with a longer lifetime, which are not already cached by a nearby node, thus improving the caching diversity in the neighborhood. As a result, vehicles can find the majority of distinct fresh and popular contents nearby, without flooding the network with content requests that have to reach the original source. Performance evaluation shows that the conceived solution outperforms representative benchmark schemes, by guaranteeing, among others, the shortest content retrieval time and the lowest network traffic load.

Automated summary

In this study, a novel distributed caching strategy is proposed, where vehicles autonomously decide which content to cache locally to enhance caching diversity and accelerate content retrieval, ultimately achieving better delivery performance in vehicular networks.