|
Activity Number:
|
115
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 9, 2004 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Physical and Engineering Sciences
|
| Abstract - #301886 |
|
Title:
|
Wavelet Methods for the Detection of Anomalies and Their Application to Network Traffic Analysis
|
|
Author(s):
|
Deukwoo Kwon*+ and Marina Vannucci
|
|
Companies:
|
Texas A&M University and Texas A&M University
|
|
Address:
|
3143 TAMU, College Station, TX, 77843-3143,
|
|
Keywords:
|
network traffic anomalies ; online detection ; wavelet packet transform ; iterated cumulative sum of squares ; Schwarz information ; changepoint detection
|
|
Abstract:
|
We develop an integrated tool for the on-line detection of network anomalies. In order for the new approach we consider two kinds of changepoint detection algorithms. One is for the variance change detrection the other the jump detection. For the former detection we modify iterative cumulative sum of squares (ICSS) and the test based on the Schwarz information criterion (SIC). For the latter we also modify the jump detection suggested by Wang (1995). Those are all for the off-line detection methods. We modify those algorithms with moving window technique for the online detection purpose. We make simulated network traffic data with several sophisticated types of attacks against the network. We mainly use the wavelet transform since wavelet transform allows us to be able to use those algorithms. We also examine the performance of the ICSS and the SIC with simulated data and compare two algorithms using the mean delay concept.
|
