Development of Education and Training Resources for GENI Experimenters

Principal Investigator: Kevin Jeffay

Funding Agency: Rayetheon, National Science Foundation

Agency Number: 14059

Abstract:

The goal of this project is to port a suite of tool for realistic synthetic traffic generation to GENI clusters (testbeds) and to develop training materials using these tools to teach prospective (and current) GENI researchers how to perform prototypical experiments on GENI clusters (and combinations of clusters). For the tools we will develop, the emphasis will be on demonstrating methods that enable GENI experimenters to generate realistic and reproducible synthetic traffic in their experiments. Our tools will allow researchers to test network elements, both hardware elements such as switches and routers, as well as software elements such as routing protocols, queue and traffic management disciplines, etc., under provably realistic conditions. In addition, our tools will enable researchers building services and protocols, to test their systems in the presence of realistic background traffic (for example to measure the effects of the application or service on the background traffic), or to use our tools to emulate applications running on top of their service or protocol. For training materials, we will develop “laboratory manuals” for GENI clusters that describe in detail the design and realization of canonical experiments for each of the current GENI clusters. Beyond serving as a “howto” manual for GENI clusters, the manuals will teach best practices methods of experimentation and include topics such as generation of synthetic and background traffic, network measurement, and data analysis. Upon completion of this project the PI will commit to combining the manuals produced in this project into an advanced undergraduate/graduate level textbook on methods and practices for experimental networking with GENI experiments serving as an extensive case study.