Edge Caching Architecture for Media Delivery over P2P Networks

Abstract

Peer to Peer (P2P) systems, have been identified as one of the main contributors to the exponential growth of internet traffic. File sharing applications such as BitTorrent clients, are the most popular among P2P systems and tend to consume great amounts of network resources, stretching infrastructure capabilities to their limits. This causes interdomain links to be overloaded with traffic, created from content traversing the network, which is most of the times redundant. Cloud computing has been an empowering force towards alleviating this issue and many research approaches have been proposed. With cloud orchestration frameworks and Software Defined Networks (SDN) as our empowering technologies, we propose a cross-layered edge-caching mechanism to alleviate content and service providers of link-saturation problems, caused by identical content requested and downloaded by multiple users in the same network vicinity. The proposed mechanism identifies and classifies P2P data-in-transit, caches it locally and successively acts as a peer by sharing it with all other requesting users. This way, redundant traffic in the backhaul link is minimized, since requests for content cached by the mechanism, will be served by the target edge-node and not from the content provider. To validate the performance and overall efficacy of the proposed mechanism, simulation experiments were conducted under a controlled-conditions environment. Results indicated that the proposed mechanism can reduce redundant network traffic up to 22%.