Jay Aikat, University of North Carolina at Chapel Hill, firstname.lastname@example.org (your questions are welcome!)
Kevin Jeffay, University of North Carolina at Chapel Hill
For over a decade now, networking researchers have built small and large research testbeds using hardware in their laboratories, or using software tools like the Network Simulator (NS) as the primary means for performance evaluation experiments. More recently, through the largely NSF-funded Global Environment for Network Innovations (GENI) project, we have witnessed the initial deployment of customizable testbeds with vast CyberInfrastructure capabilities spread across the United States.
GENI is a virtual laboratory at the frontiers of network science and engineering, enabling researchers to build and tear down networks and run experiments with several options for testbed topologies, instrumentation and measurement, and models for traffic generation. Through this tutorial, we would like to introduce the SIGMETRICS / Performance community to the vast testbeds, tools and resources openly available through the GENI project. This project, while still being developed and deployed, is now ready for running experiments.
This tutorial will answer the following questions:
o What is GENI?
o What are the distributed computing resources available on GENI for researchers interested in simulation as well as measurement-based performance evaluation experiments?
o Where and how do you get started if you want to run experiments on GENI testbeds?
o What are the kinds of experiments you can run on the GENI testbeds?
o What are the specific traffic generation tools available for such experiments?
We will demonstrate simple experiments that can be run using the GENI CyberInfrastructure. We will have a hands-on tutorial component.
** Come prepared with your laptops! **
We will use the ProtoGENI testbeds using Flack.
Using a case study of experiments for network performance evaluation, we shall walk you through the entire process from measurement of production network traffic to modeling workloads and end-to-end paths to generating this traffic for experiments using the GENI infrastructure.
This tutorial is designed for any (rookie or experienced) network researcher or network engineer who is interested in using the GENI CyberInfrastructure for their experiments. You may be a:
o Researcher who wishes to run network experiments modeling realistic traffic workloads
o Professor who wishes to add simple pre-tested GENI experiment modules into their current networking courses.
o Industry researcher or engineer who runs experiments to test and evaluate distributed systems or network components
o Most importantly, you may wish to run such experiments with very little effort at setting up the testbeds and developing tools for running experiments; i.e. GENI testbeds are available for running experiments and we will introduce you to the tools and manuals to use them.
Assumed Background of Attendees:
We expect the attendee to have basic knowledge of computer networking and distributed systems, and to be interested in the research and education of empirical networking.
o Jay Aikat, University of North Carolina at Chapel Hill, email@example.com
o Kevin Jeffay, University of North Carolina at Chapel Hill, firstname.lastname@example.org
Jay Aikat is a Research Assistant Professor in the Department of Computer Science at the University of North Carolina at Chapel Hill. Her interests include computer networking, measurement and modeling of Internet traffic, and experimental design and evaluation.
Kevin Jeffay is the Gillian T. Cell Distinguished Professor of Computer Science in the Department of Computer Science at the University of North Carolina at Chapel Hill. His research interests include computer networking, operating systems, real-time systems, multimedia networking, and performance evaluation.
The presenters of this tutorial are currently working on a three-year NSF-funded project to develop education and training resources for GENI experiments.
1. Felix Hernandez-Campos. Generation and Validation of Empirically-Derived TCP Application Workloads. PhD thesis, University of North Carolina at Chapel Hill, August 2006.
2. Michele C. Weigle, Prashanth Adurthi, Felix Hernandez-Campos, Kevin Jeffay, and F. Donelson Smith. A tool for generating realistic TCP application workloads in ns-2. ACM Computer Communication Review, 36(3):67-76, July 2006.
3. GENI: www.geni.net