The 1999 DARPA off-line intrusion detection evaluation

Citation
R. Lippmann et al., The 1999 DARPA off-line intrusion detection evaluation, COMPUT NET, 34(4), 2000, pp. 579-595
Citations number
24
Categorie Soggetti
Information Tecnology & Communication Systems
Journal title
COMPUTER NETWORKS-THE INTERNATIONAL JOURNAL OF COMPUTER AND TELECOMMUNICATIONS NETWORKING
ISSN journal
13891286 → ACNP
Volume
34
Issue
4
Year of publication
2000
Pages
579 - 595
Database
ISI
SICI code
1389-1286(200010)34:4<579:T1DOID>2.0.ZU;2-F
Abstract
Eight sites participated in the second Defense Advanced Research Projects A gency (DARPA) off-line intrusion detection evaluation in 1999. A test bed g enerated live background traffic similar to that on a government site conta ining hundreds of users on thousands of hosts. More than 200 instances of 5 8 attack types were launched against victim UNIX and Windows NT hosts in th ree weeks of training data and two weeks of test data. False-alarm rates we re low (less than 10 per day). The best detection was provided by network-b ased systems for old probe and old denial-of-service (DoS) attacks and by h ost-based systems for Solaris user-to-root (U2R) attacks. The best overall performance would have been provided by a combined system that used both ho st- and network-based intrusion detection. Detection accuracy was Door for previously unseen, new, stealthy and Windows NT attacks. Ten of the 58 atta ck types were completely missed by all systems. Systems missed attacks beca use signatures for old attacks did not generalize to new attacks, auditing was not available on all hosts, and protocols and TCP services were not ana lyzed at all or to the depth required. Promising capabilities were demonstr ated by host-based systems, anomaly detection systems and a system that per forms forensic analysis on file system data. (C) 2000 Elsevier Science B.V. All rights reserved.