新闻中心

【学术沙龙】Large scale BGP monitoring: a big data and machine learning perspective

发布时间:2018-07-19 09:36:40 发布者:研究生管理办公室 作者: 查看:1125

时间:2018年7月19日下午17:00—18:00

地点:沙河校区主楼中216会议室

主办:信息与软件工程学院

范围:全校(可盖学术章)

 

报告主题:Large scale BGP monitoring: a big data and machine learning perspective 

摘要: The monitoring of large dynamic networks is a major challenge for a wide range of application. The complexity stems from properties of the underlying graphs, in which slight local changes can lead to sizable variations of global properties, e.g., under certain conditions, a single link cut that may be overlooked during monitoring can result in splitting the graph into two disconnected components. Moreover, it is often difficult to determine whether a change will propagate globally or remain local.. In this presentation , we tackle the problem of real-time monitoring of dynamic large scale graphs by developing a geometric approach that leverages notions of geometric curvature and recent development in graph embeddings using Ollivier-Ricci curvature. We illustrate the use of our method by considering the practical case of monitoring dynamic variations of global Internet using topology changes information provided by combining several BGP feeds. In particular, we use our method to detect major events and changes via the geometry of the embedding of the graph. . We develop an anomaly tracking mechanism to detect large variations of curvature that translate into major variations in the geometry of the graph. These changes are then considered as major BGP level events. In a second stage, we describe a mechanism for identifying the network elements responsible for the set of coordinated changes and isolate their geometric implications. We evaluate this system in operational settings and show its performance in real-life scenarios.

主讲人简介: Kavé Salamatian is a full professor of computer science  at University of Savoie. His main areas of researches has been Internet measurement and modeling, network security, and networking information theory. He was previously reader at Lancaster University, UK and associate professor at University Pierre et Marie Curie. Kavé has graduated in 1998 from Paris SUD-Orsay university where he worked on joint source channel coding applied to multimedia transmission over Internet for his Phd. In a former life, he graduatedwith a MBA, and worked on market floor as a risk analyst and enjoyed being an urban traffic modeler for some years. He is currently distinguished visiting professor at the Chines Academy of Science and also working closely with the Castex CyberStrategy Chair at the French National Defense Institute. He is working these day on figuring out if networking is a science or just a hobby and if it is a science what are its fundamentals. In particular, he has been working on cyber-strategy and cyber-geography. He has published more than 150 papers.