An Evaluation Framework for Intrusion Detection Dataset

TLDR

The increasing prevalence of internet security threats and the critical role of IDS/IPS systems highlight the need for reliable benchmark datasets, yet few studies have examined the datasets themselves. This study aims to comprehensively evaluate existing IDS/IPS datasets using proposed criteria and to propose an evaluation framework. The authors assess current datasets against their criteria and design a framework to guide future dataset evaluation.

Abstract

The growing number of security threats on the Internet and computer networks demands highly reliable security solutions. Meanwhile, Intrusion Detection (IDSs) and Intrusion Prevention Systems (IPSs) have an important role in the design and development of a robust network infrastructure that can defend computer networks by detecting and blocking a variety of attacks. Reliable benchmark datasets are critical to test and evaluate the performance of a detection system. There exist a number of such datasets, for example, DARPA98, KDD99, ISC2012, and ADFA13 that have been used by the researchers to evaluate the performance of their intrusion detection and prevention approaches. However, not enough research has focused on the evaluation and assessment of the datasets themselves. In this paper we present a comprehensive evaluation of the existing datasets using our proposed criteria, and propose an evaluation framework for IDS and IPS datasets.