PFECC: a precise feedback-based explicit congestion control algorithm in named data networking

Abstract

Named data networking (NDN), as a future Internet architecture that is a promising alternative to TCP/IP networks, has the new features of connectionless, in-network cache, and hop-by-hop forwarding, which makes the congestion control algorithms of traditional TCP/IP networks unable to be directly applied to NDN. In addition, since the optimal size of the sending window cannot be determined, the existing window-based congestion control algorithms generally use the AIMD-like window adjustment algorithm, which cannot achieve the optimal throughput. In this paper, we propose a precise feedback-based, multipath-aware congestion control algorithm PFECC, which is inspired by Accel-Brake Control algorithm. PFECC considers the influence of Interest flows, uses a fair queuing algorithm at the intermediate node to calculate the target rate of each flow, and gives accurate feedback on each dequeued data packet. The consumer changes the size of the sending window according to the feedback to quickly converge to the target bandwidth. To fully exploit the hop-by-hop adaptive forwarding feature of NDN, each downstream node timely senses the congestion trend of the upstream link to split the forwarding rate of Interest to avoid congestion. Simulation results show that PFECC can effectively reduce transmission delay, treat each flow fairly, and converge to the best throughput faster.