命名数据网络中基于自适应转发的拥塞控制机制

CONGESTION CONTROL MECHANISM BASED ON ADAPTIVE FORWARDING OVER NAMING DATA NETWORKING

为了提高命名数据网络NDN(Named Data Networking)中视频数据的可靠传输,站在客户端的角度,提出一种基于自适应转发的拥塞控制机制AFCCP(Adaptive Forward Congestion Control Policy)。AFCCF以网络丢包最小化为目标,为接口的选择建立马尔科夫模型,通过前一时间间隔链路的状态,自适应地选择最佳的转发接口,减少兴趣包向拥塞链路的转发,降低网内的丢包数目,实现网络拥塞控制。在此基础上,AFCCF针对视频数据内部属性,考虑解码端特点,在网络发生丢包时,选择性地对数据包丢弃,实现视频内部重要数据的可靠传输。仿真结果表明,AFCCP在较低时延的条件下,实现网络较小的丢包率,增加用户接收数据包的数量,从而改善用户获取数据体验。

To improve the reliable transmission of video data in the named data network, this paper proposes an Adaptive Forward Congestion Control Policy （AFCCP）from the client＇ s point of view. AFCCF aims at minimizing network packet loss and establishes a Markov model for the selection of interfaces. AFCCF adaptively selects the best forwarding interface through the state of the previous time interval link and to reduce the forwarding of interest packets to the congested link. Reducing the number of packet loss in the network to realize network congestion control. On the basis of this, AFCCF aims at the internal properties of video data, consider the characteristics of the decoding side when network occurs packet loss, AFCCP will selectively discard packet, to achieve reliable transmission for the important data within the video. The simulator results show that AFCCP achieves a low packet loss rate at low latency and increases the number of packets received by the user, thus improving the user＇ s access to data experience.