阅读报告_A [2017][IFIP]On Broadcast-based Self-Learning in Named Data Networ

论文题目

On Broadcast-based Self-Learning in Named Data Networking

论文作者、单位

Junxiao Shi, Eric Newberry, Beichuan Zhang ——The University of Arizona

发表的会议或者期刊

2017 IFIP C类会议

论文关键字

略

论文摘要原文

In local area networks and mobile ad-hoc networks,broadcast-based self-learning is a common mechanism to findpacket delivery paths. Self-learning broadcasts the first packet,observes where the returning packet comes from, then createsthe corresponding forwarding table entry so that future packetswill only need unicast. The main benefits of this mechanism areits simplicity, adaptability, and support of mobility. While thehigh-level idea of broadcast-based self-learning is straightforward,making the scheme efficient and secure, especially in adata-centric network architecture like Named Data Networking(NDN), requires careful examination. In this paper, we study howbroadcast-based self-learning may be applied to NDN networks,point out two major issues: the name-prefix granularity problemand the trust problem, and propose corresponding solutions. Wealso apply self-learning to switched Ethernet as an example todevelop a specific design that can build forwarding tables withoutany control protocol, recover quickly from link failures, and makeuse of off-path caches. Simulations are conducted using both realand synthetic traffic to evaluate the performance of the design.

论文摘要中文

在局域网、移动自组织网络中，基于广播的“自学习”是一种常见的找到包传递路线的方法。自学习方法广播第一个包，观察包从哪里回来，然后创建对应的转发表项，使得将来的包只需要单播即可。该方法的主要好处是简洁性、适用性、支持移动性。虽然基于广播的自学习的高层次思想很简单，但是为了使该方法更高效率、安全，尤其是在类似NDN的数据中心网络架构中，该方法仍需进一步检验。在本文中，我们研究了基于广播的自学习如何被应用于NND，指出了两个主要问题：名称前缀粒度问题、信任问题，并给出了对应的解决方案。
作为一个例子，我们还将自学习应用于交换式以太网，开发了一个具体设计——可以在没有任何控制协议的情况下构建转发表、快速从链路失效中恢复、利用其他线路的缓存。仿真同时使用了真实、合成的流量去评估设计的性能。

研究问题

介绍了自学习广播法。着重介绍了两个值得关注的地方——(1) FIB prefix的粒度问题; (2) 认证;
将该方法应用于以太交换局域网，介绍了如何构建转发表、从链路失效中快速恢复、利用缓存。

研究意义

Broadcast-based self-learning allows networks to adapt to changing environments and allows for producer mobility, without using any routing or other control protocols.
particularly notable in mobile ad hoc and wireless network environments.
This mechanism also fits the NDN architecture well because it does not require prior knowledge of the location of data.