A new arXiv preprint introduces GreenNIDS, a benchmark workflow and analysis dataset for flow-based intrusion detection. The authors note that comparing experiments fairly requires both predictive performance and computational cost to be available for every run. GreenNIDS addresses this by connecting benchmark execution on public NetFlow datasets to a structured collection of results.
Because the source is a single preprint, there are no conflicting findings to reconcile. The paper's contribution is methodological: it offers a way to standardize how intrusion-detection experiments are measured and compared. This could help researchers move beyond accuracy-only reporting and account for the resources required to train and run models.
The abstract does not specify the exact contents of the dataset or the metrics included, but the stated goal is clear. By tying execution to a shared analysis dataset, GreenNIDS aims to make experiments more reproducible and comparisons more meaningful across different detection approaches.