Experimental Studies Using SOARA: An Approach to Reduce Alarm Rates on Streams of Intrusion
Jorge Levera (1), Robert Grossman (1), Benjamín Barán (2)
e-mails: jlevera@cs.uic.edu, grossman@uic.edu, bbaran@cnc.una.py
(1) University of Illinois at Chicago - Departament of Computer Science Chicago Estados Unidos
(2) Universidad Nacional de Asunción - Centro Nacional de Computación San Lorenzo Paraguay
The overwhelming number of alarms generated by rule-based
network intrusion detection systems makes the task of
network security operators ineffective. Preliminary results
on an approach called SOARA shows that false positive alarms
can be reduced by detecting changes on streams of alarms
using sketch-based time-decaying moving median. SOARA keeps
a memory efficient sketch summary of the normal stream of
alarms using relevant features. Sketches are updated
according to established policies and a time-decaying moving
median procedure is used on historical data to detect
abnormal alarm rates on the stream. SOARA shows promising
results on labeled and unlabeled test sets by focusing on
exceptions on the normal stream of alarms, diverting the
attention away from false
positives.
Keywords:Data Stream, Intrusion Detection, Sketch Summaries, Time Decaying Moving Median
BibTex
@INPROCEEDINGS{levera04:156,
AUTHOR = {Jorge Levera and Robert Grossman and Benjamín Barán},
TITLE = {Experimental Studies Using SOARA: An Approach to Reduce Alarm Rates on Streams of Intrusion},
BOOKTITLE = {30ma Conferencia Latinoamericana de Informática (CLEI2004)},
YEAR = {2004},
editor = {Mauricio Solar and David Fernández-Baca and Ernesto Cuadros-Vargas},
pages = {512--522},
address = {},
month = Sep,
organization = {Sociedad Peruana de Computación},
note = {ISBN 9972-9876-2-0},
file = {http://clei2004.spc.org.pe/es/html/pdfs/156.pdf}
}
PDF de este artículo
PDF de CLEI2004 (incluye todos los artículos)
Página principal CLEI 2004
Generado por Sociedad Peruana de Computación