arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1202.3993cs.NIcs.SI
收藏

随时间变化的互联网拓扑

Internet Topology over Time

Benjamin Edwards, Steven Hofmeyr, George Stelle, Stephanie Forrest

摘要:

很少有研究仔细研究互联网的拓扑结构如何随着时间的推移而演变;大多数研究都集中在特定时间点拍摄的快照上。在本文中,我们研究了互联网的自治系统图的拓扑结构的演变,研究如何8个常用的拓扑措施从2002年1月到2010年1月的变化。我们发现,除了平均路径长度和聚类系数外,大多数测度的分布保持不变。自2005年以来,平均路径长度缓慢而稳定地增加,平均聚类系数稳步下降。我们假设这些变化是由于互联网发展的对等策略的变化。我们还调查了一个令人惊讶的功能,即最大程度的变化不大,一个方面,不能捕捉到没有建模链接删除。我们的结果表明,通过将稳态生成的拓扑与真实数据的快照进行比较来评估互联网图的模型对于许多措施来说都是合理的。然而,准确地匹配时变属性是更困难的,因为我们证明了对2010年的数据评估10个著名的模型。

英文摘要:

There are few studies that look closely at how the topology of the Internet evolves over time; most focus on snapshots taken at a particular point in time. In this paper, we investigate the evolution of the topology of the Autonomous Systems graph of the Internet, examining how eight commonly-used topological measures change from January 2002 to January 2010. We find that the distributions of most of the measures remain unchanged, except for average path length and clustering coefficient. The average path length has slowly and steadily increased since 2005 and the average clustering coefficient has steadily declined. We hypothesize that these changes are due to changes in peering policies as the Internet evolves. We also investigate a surprising feature, namely that the maximum degree has changed little, an aspect that cannot be captured without modeling link deletion. Our results suggest that evaluating models of the Internet graph by comparing steady-state generated topologies to snapshots of the real data is reasonable for many measures. However, accurately matching time-variant properties is more difficult, as we demonstrate by evaluating ten well-known models against the 2010 data.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑