the subspace method for diagnosing network wide traffic
play

The Subspace Method for Diagnosing Network-Wide Traffic Anomalies - PowerPoint PPT Presentation

The Subspace Method for Diagnosing Network-Wide Traffic Anomalies Anukool Lakhina, Mark Crovella, Christophe Diot Whats happening in my network? Is my customer being attacked? probed? infected? Is there a sudden traffic shift?


  1. The Subspace Method for Diagnosing Network-Wide Traffic Anomalies Anukool Lakhina, Mark Crovella, Christophe Diot

  2. What’s happening in my network? • Is my customer being attacked? probed? infected? • Is there a sudden traffic shift? • An external route change? • A routing loop? • An equipment outage? Automated methods for reliably and generally answering such questions are lacking 2

  3. A General Framework • We can treat all such problems as special cases of the general question: Is my network experiencing unusual conditions? • Then, adopt the following framework: – Detection Is there an unusual event? – Identification Which of the possible explanations fits best? – Quantification How serious is the problem? 3

  4. Statistical Approach The advantage of such a framework is that it lends itself to a statistical approach: – Detection: Outlier detection Anomaly – Identification: Hypothesis testing Diagnosis – Quantification: Estimation 4

  5. A Need for Whole-Network Diagnosis Our Thesis: Effective diagnosis of network anomalies requires a whole-network approach For example, diagnosing traffic anomalies requires analyzing traffic from all links 5

  6. But, This Is Difficult! x 10 7 DNVR−SNVA x 10 7 HSTN−KSCY x 10 8 LOSA−SNVA x 10 8 CHIN−NYCM • Need to study traffic from all links in a 6 2 6 5 5 1.5 4 4 network simultaneously 5 1 3 3 4 2 2 0.5 1 3 – Large amount of data 1 x 10 7 ATLA−ATLA 3 x 10 7 DNVR−DNVR x 10 6 KSCY−KSCY x 10 7 SNVA−SNVA 4.5 5 – Traffic is nonstationary 15 2.5 4 4 2 3.5 10 3 1.5 3 – Varying link utilization levels 2 5 2.5 1 1 2 0.5 – 100s of links � High dimensionality x 10 6 LOSA−LOSA x 10 7 HSTN−HSTN x 10 6 STTL−STTL x 10 8 WASH−WASH 6 4 2 5 1.5 3 4 1.5 1 3 2 2 1 1 0.5 1 0 How do we extract meaning from such a high-dimensional data in a systematic manner? 6

  7. Low Intrinsic Dimensionality of Link Traffic Studied via Principal Component Analysis Key result: Normal traffic is well approximated by a low dimensional space For example: Traffic on 40+ links is well approximated in space of only 4 dimensions 7

  8. Reasons for Low Dimensionality of Traffic • Generally, traffic on different links is not independent • Link traffic is the superposition of origin- destination flows (OD flows) – The same OD flow passes over multiple links, inducing correlation among links – All OD flows tend to vary according to common daily and weekly cycles, and so are themselves correlated [See SIGMETRICS 2004 paper] 8

  9. The Subspace Method • An approach to separate normal from anomalous traffic • Define as the space spanned by the first k principal components • Define as the space spanned by the remaining principal components • Then, decompose traffic on all links by projecting onto and to obtain: Residual traffic Traffic vector of all Normal traffic vector links at a particular vector point in time 9

  10. The Subspace Method, Geometrically In general, anomalous traffic results in Traffic on Link 2 a large value of y Traffic on Link 1 10

  11. Outline • Subspace Method applied to Link Traffic – Problem: Volume Anomaly Diagnosis – Detection, Identification, Quantification – Validation • Subspace Method applied to Flow Traffic – Problem: General Anomaly Detection – Sample Results • Conclusions 11

  12. Diagnosing Volume Anomalies • A volume anomaly is a sudden change in an OD flow’s traffic ( i.e., point to point traffic) • Problem Statement: Given link traffic measurements, diagnose the volume anomalies • A first application of the subspace method 12

  13. An Illustration 6 x 10 OD flow i−b 15 10 5 7 x 10 Link c−b 8 Sprint-Europe Backbone Network 6 4 7 x 10 Link d−c The Diagnosis Problem requires 6 4 analyzing traffic on all links to: 2 7 x 10 Link f−d 1) Detect the time of the anomaly 6 4 2) Identify the source & destination 2 7 x 10 Link i−f 6 3) Quantify the size of the anomaly 4 2 Fri Sat Sun 13

  14. Subspace Method: Detection • Error Bounds on Squared Prediction Error: • Assuming multivariate Gaussian data, traffic is Traffic on Link 2 normal when, Result due to [Jackson and Mudholkar, 1979] Traffic on Link 1 14

  15. SPE vs. All Traffic Value of over time Value of over time SPE ( ) at anomaly time points clearly stand out 15

  16. Results on True Anomalies: Sprint-1 40 Largest deviations in OD flows via Fourier Detection Quantification Identification “Knee” in curve - natural cutoff for detection 16

  17. Outline • Subspace Method applied to Link Traffic – Problem: Volume Anomaly Diagnosis – Detection, Identification, Quantification – Validation • Subspace Method applied to OD Flow Traffic – Problem: General Anomaly Detection – Sample Results • Conclusions 17

  18. Beyond Volume Anomalies • Volume anomalies: important, but not the entire set of anomalies of interest to operators. • Operators are also interested in: – DOS attacks, flash crowds, port scans, worm propagation, network equipment outages, changes in ingress/egress traffic patterns, ... • Link data doesn't seem to hold enough information to accurately detect such a wide range of anomaly types. • Therefore, we turn to IP flow data 18

  19. Characterization Methodology • Extend subspace method to diagnose anomalies directly in OD flow traffic timeseries – Detection in both and subspaces • Examine OD flow traffic as three separate views: # Bytes, # Packets, # IP-flows • Manually inspect each anomaly found over 4 week period in Abilene network – Using 5-tuple headers of sampled flow data 19

  20. An example BP anomaly (heavy flow) Dominant Source IP: 192.88.112.0 which accounts for 32% of B, 20% of P and 0.15% of F . Dominant Dest. IP: 160.91.192.0 which accounts for 32% of B, 20% of P and 0.15% of F. Dominant Pair: 192.88.112.0-160.91.192.0 for 32% of B, 20% of P and 0.15% of F. Dominant Dest. Port: 5002 (iperf port, used by SLAC) 20

  21. An example PF anomaly (DOS attack) Dominant Source IP: No dominant single source Dominant Dest. IP: 211.65.112.0 accounts for 80% of P traffic and 92% of F traffic. Dominant Pair : No single pair dominant Dominant Ports : No dominant source or destination port found 21

  22. An example BPF Anomaly (ingress-shift) Multihomed customer CALREN reroutes around the LOSA-CHIN (scheduled) outage 22

  23. Species of anomalies found Anomaly Definition ALPHA Unusually high rate point to point byte transfer DOS, DDOS (Distributed) Denial of service attack against a single victim FLASH CROWD Unusually large demand for a resource/service emerging from common set of sources SCAN Scanning a host for a vulnerable port (port scan) or scanning the network for a target port (network scan) WORM Self-propagating code that spreads across a network by exploiting security flaws POINT to Distribution of content from one server to many servers MULTIPOINT OUTAGE Equipment related events that decrease traffic exchanged by an OD pair INGRESS-SHIFT Customer shifts traffic from one ingress point to another 23

  24. Summary of Anomalies Found 31 39 137 4 Alpha 3 DOS 2 Scan 3 Flash−Crowd Point−Multi Worm Outage Ingress−Shift Unknown FalseAlarm 64 24 44 56

  25. Conclusions • Subspace method for anomaly diagnosis allows whole- network approach – Significant benefit accrues from whole-network analysis • Diagnosing Volume Anomalies from Link Traffic: – High detection rate, low false alarm rate – Hypothesis-based identification is easily formalized and extended • Detecting General Anomalies from Flow Traffic: – Anomalies detected span remarkable breadth – Almost all of the anomalies found are operationally relevant • Whole-Network Anomaly Diagnosis with the Subspace Method is promising – ... more to come! 25

  26. Thanks! Help with Abilene Data • Rick Summerhill, Mark Fullmer (Internet2) • Matthew Davy (Indiana University) Help with Sprint-Europe Data • Bjorn Carlsson, Jeff Loughridge (SprintLink), • Supratik Bhattacharyya, Richard Gass (ATL) 26

Recommend


More recommend