Combing CCN with network coding: An architectural perspective
Combing CCN with network coding: An architectural perspective
复制标题
DOI：
10.1016/j.comnet.2015.11.008
复制
发表时间：
2016-01
期刊：
影响因子：
--
通讯作者：
Guoqiang Zhang;Ziqu Xu
中科院分区：
文献类型：
--
作者：
Guoqiang Zhang;Ziqu Xu
以内容为中心的网络和网络编码都围绕着内容分发，分别为带宽交换存储和计算。尽管人们认为网络编码有利于CCN，但真正的优势、架构含义和设计选择并没有得到充分的探讨。在本文中，我们首先研究了将网络编码与CCN结合的可能优势。我们认为这样做可以消除CCN中的细粒度命名，有效地利用多路径并行转发，降低缓存协调的复杂性，简化传输设计。然后，我们分析了将网络编码与CCN结合的体系结构含义和设计选择，特别是由于命名的语义差异而引起的体系结构变化。最后，根据架构含义，我们提出了一个称为NC-CCN的原型实现，并通过大量模拟评估其性能。我们比较了两种线性相关性检查的匹配机制，发现基于秩的匹配机制（RB-Matching）的性能略低于精确匹配机制（precision - matching），但计算和通信开销更低。
Content-centric networking and network coding both revolve around content distribution, trading storage and computation for bandwidth respectively. Although it was believed network coding benefits CCN, the real advantages, architectural implications and design choices are not fully explored. In this paper, we first investigated the possible advantages of combing network coding with CCN. We argue doing so can eliminate the fine-grained naming in CCN, effectively utilize multi-path parallel forwarding, reduce the complexity of cache coordination and simplify the transport design. We then analyzed the architectural implications and design choices for incorporating network coding with CCN, especially the architectural changes arising from the semantic difference of naming. Finally, following the architectural implications, we proposed a prototype implementation called NC-CCN and evaluated its performance by substantial simulations. We compared two matching mechanisms for linear dependency checking, and found that the rank-based matching mechanism(RB-Matching) can achieve slightly lower performance compared with precise matching mechanism(Precise-Matching) with much lower computation and communication overhead.