STNDN: Link Aware Segmented Transmission for Named Data Networking
Jin Qin,Kaiping Xue,Yutong Chen,Wenjia Wei,Jianqing Liu,Hao Yue
DOI: https://doi.org/10.1109/HotICN48464.2019.9063211
2019-01-01
Abstract:Named data networking (NDN) is a promising implementation of information centric networking (ICN), which is based on a publish/subscribe model. In NDN, when an interest packet containing the name of requested content is sent to network, the corresponding data chunk will be returned from a content server or an intermediate router that stores a copy of the content. In order to avoid interest flooding, chunk size usually cannot be set too small and larger chunk will be divided into multiple smaller fragments in transmission. However, in existing NDN designs, the fragmentation, aggregation, and acknowledgment of chunks need to be performed at each hop, which brings huge processing delay. In this paper, we propose a segmented transmission scheme for NDN (STNDN) to address this problem. In STNDN, each fragment of chunk can be independently routed, and chunks will only be aggregated at some selected nodes
<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">a</sup>
long the transmission path. We also consider the impact of link state instability. Specifically, w e a ggregate c hunks a t the u pstream of each poor link to avoid retransmissions across multiple hops on these links, thus improving the robustness of STNDN. We further demonstrate the effectiveness of STNDN through experiments in ndnSIM simulator. Compared with other two solutions, NDNLP and FFRD, STNDN has significantly improved both the reliability and efficiency of transmission.