Forum Numerica - Walid Dabbous "Towards Reproducible Wireless Networking Experiments" Sept. 27, 2018
Reproducibility is key in designing wireless systems and evaluating their performance. Trying to reproduce wireless experiments allowed us to identify some pitfalls, and possible ways to simplify the complex task of avoiding them. In this talk we expose a few considerations that we learned are instrumental for ensuring reproducibility of wireless experiments. We specifically address issues related to wireless hardware, as well as varying propagation channel conditions. We show that an extensive knowledge of the used hardware and of its design is required to guarantee that the inner state of the system has no negative impact on performance evaluation and experimental results. As for variability of channel conditions, we make the case that a special setup or testbed is necessary so that one can control ambient wireless propagation environment, using for instance an anechoic chamber like R2lab. In order to ease the task of repeating and modifying the experiments scenarios we used the nepi-ng experiment control tool. We will explain very briefly this tools’ features and demonstrate its use through a simple Python notebook. This talk is prepared to be suitable for a large audience.
Walid Dabbous is a Senior Researcher with Inria, where he is a Leader of DIANA Team. His research interests include networking experimental platforms and simulators, and experimental methodology for evaluating networking protocols.
FORUM NUMERICA is sponsored by the UCA JEDI “Networks, Information and Digital Society” Academy.