“An Evaluation of Effect of Packet Sampling on Anomaly Detection Method” Takuya Motodate April 25, 2010 The 3rd CAIDA-WIDE-CASFI Joint Measurement Workshop @Osaka 1 2010 年 4 月 25 日日曜日
Background • Anomaly Detection: Signature-based,Statistical one • Statistical anomaly detection assumes a full-captured dump. • Traffic of backbone network become broader, so, characteristics of it is grasped with sampled traffic. • We have to use sampled-traffic as input of anomaly detection. What should we do? 2 2010 年 4 月 25 日日曜日
Problem Statement 1. Suitable Packet-Sampling Method is not Known. 2. Suitable Anomaly Detection Method is not Known. Because of inadequate evaluations. 3 2010 年 4 月 25 日日曜日
Purpose • Evaluate an effect to result of anomaly detection methods with various sampling methods and common traffic data. • Sketch and Non Gaussian Multi-Resolution Statistical Detection Procedures as a Anomaly Detection Method. • 5 Packet-Sampling Methods. • MAWI Dataset as Traffic Data. 4 2010 年 4 月 25 日日曜日
1.Divide a traffic into some subtraffics. 2. Estimate α and β of each subtraffic, each timescale. 3. Anomalous subtraffic has deviate α or β. Sketch and Non Gaussian Multi-Resolution Statistical Detection Procedures [Dewaele et al. 07] 5 2010 年 4 月 25 日日曜日
(1)Hashing: Key is SrcIPAddr 1.Divide a traffic into some subtraffics. 2. Estimate α and β of each subtraffic, each timescale. 3. Anomalous subtraffic has deviate α or β. Sketch and Non Gaussian Multi-Resolution Statistical Detection Procedures [Dewaele et al. 07] 5 2010 年 4 月 25 日日曜日
(1)Hashing: Key is SrcIPAddr (2) Making Histgram, and 20ms 5ms 80ms 1.Divide a traffic into some subtraffics. 2. Estimate α and β of each subtraffic, each timescale. 3. Anomalous subtraffic has deviate α or β. Sketch and Non Gaussian Multi-Resolution Statistical Detection Procedures [Dewaele et al. 07] 5 2010 年 4 月 25 日日曜日
5ms 1.Divide a traffic into some subtraffics. 3. Anomalous subtraffic has deviate α or β. (1)Hashing: Key is SrcIPAddr (2) Making Histgram, and 20ms 5ms 80ms 2. Estimate α and β of each subtraffic, each timescale. 20ms 80ms Estimating Parameters of Gamma Distribution Sketch and Non Gaussian Multi-Resolution Statistical Detection Procedures [Dewaele et al. 07] 5 2010 年 4 月 25 日日曜日
80ms 5ms (3)Detecting anomaly of Gamma Distribution Estimating Parameters 2. Estimate α and β of each subtraffic, each timescale. 20ms 80ms 1.Divide a traffic into some subtraffics. 5ms 20ms (2) Making Histgram, and SrcIPAddr (1)Hashing: Key is 3. Anomalous subtraffic has deviate α or β. Anomalies Sketch and Non Gaussian Multi-Resolution Statistical Detection Procedures [Dewaele et al. 07] 5 2010 年 4 月 25 日日曜日
80ms 5ms (3)Detecting anomaly of Gamma Distribution Estimating Parameters 2. Estimate α and β of each subtraffic, each timescale. 20ms 80ms 1.Divide a traffic into some subtraffics. 5ms 20ms (2) Making Histgram, and SrcIPAddr (1)Hashing: Key is 3. Anomalous subtraffic has deviate α or β. Anomalies Sketch and Non Gaussian Multi-Resolution Statistical Detection Procedures [Dewaele et al. 07] 5 2010 年 4 月 25 日日曜日
Packet-Sampling Methodologies [Claffy et al. 93] Stratified Simple Systematic Random Random Packet-based Packet-based Packet-based Systematic Stratified Random Simple Random Random Time-based Time-based Stratified Time-based Systematic Random Packet-based : A. Systematic Sampling Picking up a packet per N packets B. Strati ed Random Sampling Time-based : Picking up a packet per M msec C. Simple Random Sampling Packet Bucket: Packet-based: ex. 100 packets Time-based: ex. 1 msec 6 2010 年 4 月 25 日日曜日
Overview of Evaluation Anomaly Result Detection MAWI Traffic Data Compare Anomaly Packet Result Sampling Detection 7 2010 年 4 月 25 日日曜日
Evaluation • I apply this evaluation to MAWI Traffic Data at 4days. - A Wednesday in December from 2004 to 2007, sample-Point B or F. Dec 15, 2004 Dec 14, 2005 Dec 13, 2006 Dec 12, 2007 8 2010 年 4 月 25 日日曜日
Numbers of Detected Hosts with each Sampling-Rate Brief Observation: 1. Detected hosts decreased as sampling-rate decreased. 2. Rapid increase is observed 2004 with time-based sampling. 9 2010 年 4 月 25 日日曜日
Parameter Tuning Target Hosts Target Hosts after Parameter Tuning Trying to make target hosts fixed. 10 2010 年 4 月 25 日日曜日
Numbers of Detected Hosts with each Sampling-Rate after normalization Brief Observation: 1. Different behavior between packet-based and time-based in high sampling-rate 2. Rapid Increase number of Simple-Random in low sampling- rate. 11 2010 年 4 月 25 日日曜日
Undergoing Things • Analysis a reason rapid increase of anomalies with simple-random in low sampling-rate, and difference between result with time-based and packet-based. • Cross-Validation: with Port-based Categorization. • Comparison with another Anomaly Detection Method. 12 2010 年 4 月 25 日日曜日
Summary • Necessary for an evaluation in using anomaly detection with sampled-traffic. • Evaluating a “Sketch and Non Gaussian Multi-Resolution” with 5 sampling methods. • Performance Difference between Time- based and Packet-based sampling, simple- random sampling. 13 2010 年 4 月 25 日日曜日
Fin. Thank you for Listening. 14 2010 年 4 月 25 日日曜日
Distribution of a number of arrival packets 2004/12/15(Wed) 14:00-14:15 1200 Original Packet-based Sampling Time-based Sampling Original 1000 Arrival Packets(pkt) 800 Packet-based 600 Time-based 400 200 0 900 200 300 400 500 600 700 800 0 100 Time(sec) Pakcet-based Systematic : 1/4 pkt/pkt Time-based Systematic: 1/4.4 pkt/pkt 15 2010 年 4 月 25 日日曜日
Recommend
More recommend
