Large-Scale Measurement Platforms

This tutorial introduces the audience to the complete Large-Scale Measurement of Broadband Performance Architectures and Protocols (LMAP) ecosystem. In the first part of the tutorial, we will give an introduction to the different large-scale measurement platforms that are currently deployed in the world, along with the insights that have been published based on data collected over multiple years. In the second part of the tutorial, we will introduce the participants to an OpenWrt-based Measurement Agent (MA). The participants will get practical training on how to write and schedule a measurement test, how to schedule the reporting of measurement results and later use a RESTful API to retrieve the results back for data analysis.  They will also provision a simulated MA in a virtualized environment under qemu to develop and debug their own measurement tests. In the third part of the tutorial, the participants will get introduced to the newly-released RIPE RESTful API for accessing user-defined measurements (UDM). They will dissect several types of publicly available UDMs, and learn how to provision custom measurements on the RIPE Atlas infrastructure. Towards the end, the participants will use the measurement data retrieved from the API to visualize the results using the data-driven javaScript library d3.js.

Instructor 1: Vaibhav Bajpai is a PhD student in the Computer Networks and Distributed Systems (CNDS) research group at Jacobs University Bremen, Germany. He received his M.Sc. degree in Computer Science from Jacobs University Bremen, Germany in 2012, and his Bachelor’s degree in Computer Science and Engineering from Uttar Pradesh Technical University (UPTU), India in 2009. His current research focuses on understanding the impact of network infrastructure changes using Large-Scale Measurement Platforms. Previously, he has worked on the area of network traffic analysis and in the NETCONF community as well.

Instructor 2: Nikolay Melnikov is a PhD student in the Computer Networks and Distributed Systems (CNDS) research group at Jacobs University Bremen, Germany. He received his Bachelor degree in "Electrical Engineering and Computer Science" in 2008, and a M.Sc. degree in "Communications, Systems and Electronics" in 2010. His current research focuses on understanding the behavioral patterns of the Internet users and the ways of differentiating Internet users with as little privacy intrusion as possible.

Hands-on tutorial for the Virtual Wall Infrastructure

The iMinds Virtual Wall is a testbed facility for setting up large-scale network topologies. The virtual wall facilities consist of 100 nodes interconnected via a non-blocking VLAN Ethernet switch. Each server is connected with 4 or 6 gigabit Ethernet links to the switch. The Virtual Wall nodes can be assigned different functionalities ranging from terminal, server, network node, and impairment node. As such, it is a generic experimental environment for advanced network, distributed software and service evaluation, and supports scalability research. The Virtual Wall is also the testbed facility that comprises one of the sites of both the FP7 BONFIRE and FP7 OFELIA projects.

In this tutorial, the presenters will give an overview of the functionality, the architecture and a hands-on guided tour for the participants. The tutorial will explain how large-scale network experiments (e.g., the deployment of a new management algorithm) can be carried out. First, the tutorial will discuss how complex network topologies can be created using either a graphical user interface or through scripts. Furthermore, we will discuss how standard network functions (e.g., web server, proxy, cache, router) can be loaded on a node and extended to modify your network experiment. The tutorial will also discuss how novel network algorithms can be tested under varying conditions by introducing dynamic traffic shaping and introducing impairments in the network.

Instructor 1: Steven Latré is a post-doctoral fellow at Ghent University, Belgium and iMinds. He obtained a M.Sc. degree in computer science from Ghent University, Belgium, in June 2006 and a Ph.D. in Computer Science Engineering in June 2011. His main research interests are the use of autonomic network management approaches with a special focus on Quality of Experience optimization and management of federations. He is, and has been, involved in several European projects, including FP7 IP Muse, CELTIC RUBENS, FP7 STREP ECODE, FP7 STREP OCEAN and FP7 NoE Flamingo.

Instructor 2: Jeroen Famaey is affiliated with the department of Information Technology at Ghent University and iMinds as a post-doctoral researcher. He received his M.Sc. degree in computer science from Ghent University in 2007 and a Ph.D. in Computer Science Engineering in June 2012. His research interests include multimedia service delivery, autonomic communications and federated network management. He is, and has been, involved in several European projects, including CELTIC RUBENS, FP7 STREP OCEAN and FP7 NoE Flamingo.

EmanicsLab: A European Research Network tailored to Network and Service Management

EmanicsLab is a European research network initiated in 2007 and funded from 2007-2009 by the European Network of Excellence for the Management of Internet Technologies and Complex Services (EMANICS). EmanicsLab is based on myPLC, the backend management infrastructure of PlanetLab. The network currently consists of 22 nodes at 11 sites across Europe. EmanicsLab partners use the network for research activities in the area of network and service management, including distributed flow collection and analysis systems, distributed intrusion detection systems, and distributed monitoring and accounting systems.

This tutorial will revisit the basic concepts of virtual distributed test-labs like PlanetLab or EmanicsLab and give a hands-on training about how to use them for research activities. First, the underlying idea and principles of PlanetLab are presented and an overview on its services and tools is given. Furthermore, EmanicsLab, a small-scale test-lab which is based on the same technology as PlanetLab is introduced and discussed. In particular, the monitoring and management capabilities of EmanicsLab will be shown. Finally, a set of practical exercises will be carried out based on a simple service which will be deployed on EmanicsLab.

Instructor: David Hausheer is assistant professor at the Department of Electrical Engineering and Information Technology of the Technische Universität Darmstadt since May 2011. He holds a diploma degree in electrical engineering and a Ph.D. degree in technical sciences from ETH Zurich. From 2005 - 2011 he has been employed as a senior researcher and lecturer at University of Zurich, Switzerland, while being on leave as a visiting scholar at EECS, UC Berkeley from October 2009 to April 2011. His current research topics include Peer-to-Peer Systems, Network Economics, Energy-efficient Networks and Future Internet.