Walter Willinger (NIKSUN): Internet Research with "Big (Internet) Data"

Abstract:

With peta-bytes of data that are continuously collected about various aspects of the Internet, how hard can it be to obtain an accurate picture of its traffic, its physical topology (i.e., router-level Internet) and the various virtual connectivity structures (e.g., AS-level topology, WWW, P2P networks, online social networks), or its “dark” sides and associated activities? In this course, Dr. Willinger will use the example of the Internet’s router topology to illustrate why despite all the available “big data”, we still lack a basic view of the physical Internet and how future Internet measurement studies have to change so that the much-heralded big data approach to Internet research can achieve ifs full potential. Simple exercises will help the students understand the nature of these changes.

Slides:

Video:

Part 1:

Part 2:

Mark Crovella (Boston University): Data mining for measurement data

Abstract:

In this short course Prof. Crovella will discuss matrix-analytic methods for data mining tasks on Internet data. He will focus on the use of linear models based on matrix factorization to address problems in (a) recovering missing data and (b) anomaly detection. Example implementations will be given using python. Students should bring one or more datasets consisting of numeric data in matrix form for use in a short lab exercise and make sure that their laptops have python (2.x or 3.x) with the following libraries: numpy, matplotlib, pandas, and scikit-learn.

Slides:

Video:

Part 1:

Part 2:

Bendert Zevenbergen (Oxford Internet Institute): Networked Systems Research: Ethical, Legal and Policy Issues

Abstract:

Ethical dilemmas in information control research and conducting Internet measurements have been a topic of increasing debate over the last few years. As the Internet is becoming increasingly pervasive in people’s
lives, networked systems research projects are collecting increasingly sensitive data on individuals, or affecting people in previously unknown ways. A discrepancy exists between human subject research - where there
are relatively strict and broadly-understood ethical traditions - and networked systems research where (a) the consideration of these issues is relatively new, and (b) the existing best practices from other fields don't easily translate. After a discussion on some ethical considerations, students will tackle some practical dilemmas in these fields in an interactive session.

Slides:

Video:

Panel : research directions on network measurements
Mark Crovella (BU), Christophe Diot (Technicolor), Paul Francis (MPI-SWS),
Pablo Rodriguez (Telefonica), Walter Willinger (Niksun)

Video: