A Novel Metric for the Evaluation of IDSs Effectiveness - Inria - Institut national de recherche en sciences et technologies du numérique
Communication Dans Un Congrès Année : 2014

A Novel Metric for the Evaluation of IDSs Effectiveness

Khalid Nasr
  • Fonction : Auteur
  • PersonId : 989392

Résumé

Nowadays intrusion detection system (IDS) has a considerable attention as a crucial element in network security. The question that arises is which IDS is effective for our system? The answer should inevitably take into account the evaluation of IDSs effectiveness. Dealing with this challenge, many valuable evaluation metrics have been introduced such as receiver operating characteristic (ROC) curve, Bayesian detection rate, intrusion detection capability, intrusion detection operating characteristic, cost-based metrics, etc. The benefits and drawbacks of these metrics are discussed in this paper. We subsequently propose a novel metric called intrusion detection effectiveness (EID) that manipulates the drawbacks of the existing ones, taking into account all essential and related parameters. We demonstrate the utility of EID over the previously proposed ones, and how it realizes the measurement of the actual effectiveness rather than the relative effectiveness as followed by the existing ones. EID can be used for evaluating the wired or wireless IDSs effectiveness. Additionally, we conduct experimental evaluation of two popular wireless IDSs (WIDSs), Kismet and AirSnare, to illustrate the benefits of EID.
Fichier principal
Vignette du fichier
978-3-642-55415-5_18_Chapter.pdf (1.06 Mo) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01370368 , version 1 (22-09-2016)

Licence

Identifiants

Citer

Khalid Nasr, Anas El Kalam. A Novel Metric for the Evaluation of IDSs Effectiveness. 29th IFIP International Information Security Conference (SEC), Jun 2014, Marrakech, Morocco. pp.220-233, ⟨10.1007/978-3-642-55415-5_18⟩. ⟨hal-01370368⟩
107 Consultations
101 Téléchargements

Altmetric

Partager

More