Geography and Routing in the Internet
Geography and Routing in the Internet
复制标题
互联网中的地理和路由
DOI：
--
复制
发表时间：
2018
期刊：
影响因子：
--
通讯作者：
M. E. Tozal
中科院分区：
文献类型：
--
作者：
A. Nur;M. E. Tozal
互联网是一个由数以万计的自治系统（AS）组成的网络网络。这些应用服务器以不同的形式相互连接，以实现“全球”互联网通信。在本研究中，我们研究了可见互联网的地理特征，并研究了地理与 AS 内和 AS 间路由策略之间的关系。我们表明，与端到端路径相比，入口到出口子路径具有较低的迂回性。我们的研究结果不仅证明了自治系统部署的高效骨干基础设施和路由方案，而且还表明了经济激励对采用自治系统间路径的影响。我们提出并检验了地理距离和往返延迟时间之间是否存在强相关性，以及地理距离和互联网中的跳长之间缺乏相关性。我们通过采用跨 AS (X-AS) 互联网拓扑图来研究地理距离和 AS 内路由策略之间的关系。我们的结果表明，无论地理距离是否用作路由决策中的自定义参数，超过三分之二的 AS 内子路径与最短地理距离一致。我们的研究结果为地理和互联网路由之间的关系提供了新的见解，这使得网络研究人员和从业者能够改进他们的网络基础设施，重新评估他们的路由策略，部署地理感知的网络覆盖，并开发更现实的网络模拟过程。
The Internet is a network of networks consisting of tens of thousands of Autonomous Systems (ASes). These ASes connect to each other in different forms to enable the “global” Internet communication. In this study, we investigate the geographical characteristics of the visible Internet as well as examine the relation between geography and intra-AS and inter-AS routing policies. We show that the ingress-to-egress subpaths have lower circuitousness compared to the end-to-end paths. Our findings not only demonstrate the efficient backbone infrastructures and routing schemes deployed by ASes but also show the consequences of economical incentives on the adoption of inter-AS paths. We present and examine the existence of a strong correlation between the geographical distance and round trip delay time as well as the lack of a correlation between the geographical distance and hop length in the Internet. We investigate the relation between the geographical distance and intra-AS routing policies by employing cross-AS (X-AS) Internet topology maps. Our results show that more than two thirds of the intra-AS subpaths are congruent with the shortest geographical distance whether or not geographical distance is employed as a custom parameter in routing decisions. Our results provide new insights into the relations between geography and Internet routing, which allow the network researchers and practitioners to improve their networking infrastructures, reevaluate their routing policies, deploy geography-aware network overlays, and develop more realistic network simulation processes.