Measuring and Analyzing DoS Flooding Experiments

Abstract

In running volumetric Denial-of-Service experiments on DeterLab, we have faced challenges regarding the collection and handling of experiment results without impinging on the performance of the experiments themselves. In this paper we report on our findings with respect to both the collection and handling of data from our experiments, based on challenges of both performance and scale. To instrument experiments with minimal interference with results, we use simple packet capture, in contrast with DeterLab's supported real-time collection and insertion of experiment data, which is prohibitively slow for our high-volume and many-experiment use case. The contribution of this work is a comparison of two approaches to experiment grouping and analysis driven by the example of our experimental requirements of high-volume and performance sensitivity.