Multipath congestion control in content-centric networks

Data communication across the Internet has significantly changed under the pressure of massive content delivery. Content-Centric Networking (CCN) rethinks Internet communication paradigm around named data retrieval, in contrast with the host-to-host transport model of TCP/IP. Content retrieval is natively pull-based driven by user requests, point-to-multipoint and intrinsically coupled with the availability of network storage. By leveraging the key features of CCN transport, in this paper we propose for the first time a congestion control mechanism realizing efficient multipath communication over content-centric networks. Our proposal is based on a Remote Adaptive Active Queue Management (RAAQM) at the receiver that performs a per-route control of bottleneck queues along the paths. We analyze the stability of the proposed solution and assess its performance by means of CCN packet-level simulations under random and optimal route selection.