Content discovery for information-centric networking

Abstract

The information-centric networking (ICN) adopts a content name as a network identifier and utilizes in-network storages to cache the contents. With the name-based routing and content caching, ICN can provide substantial benefits such as faster content retrieval and network traffic reduction by exploiting a nearby (cached) copy of content and reducing duplicated transmissions for the same content request. Prior researches on ICN usually rely on an opportunistic cache-hit (happen-to-meet) to utilize the in-network storages. In the happen-to-meet fashion, only the content cached on the path towards the content source can be utilized, which limits the network-wide usage of the in-network storages. To exploit cached contents better, we propose a content discovery scheme, dubbed SCAN, which can exploit nearby content copies for the efficient delivery. SCAN exchanges the cached content information among the neighbor routers using Bloom filters for the content discovery. With extensive simulations, SCAN shows better performance than a happen-to-meet cache-hit scheme in terms of average hop counts, traffic volume, and load balancing among links.