Model-Based Covert Timing Channels: Automated Modeling and Evasion Steven Gianvecchio 1, Haining Wang 1, Duminda Wijesekera 2, and Sushil Jajodia 2 1 College of William and Mary 2 George Mason University
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 2 Outline Background Covert Timing Channels Model-Based Framework Experimental Evaluation Capacity Detection Resistance Conclusion
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 3 Outline Background Covert Timing Channels Model-Based Framework Experimental Evaluation Capacity Detection Resistance Conclusion
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 4 Background Covert Channels manipulate shared resources to transfer information hide communication (or extra communication) exfiltrate sensitive data (e.g., keys, passwords)
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 5 Background Types of Covert Channels shared resource is the type covert storage channels (e.g., packet header fields) covert timing channels (e.g., packet arrival times)
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 6 Outline Background Covert Timing Channels Model-Based Framework Experimental Evaluation Capacity Detection Resistance Conclusion
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 7 Main Goals high capacity strong detection resistance Capacity – bits/time unit, not bits/symbol Covert Timing Channels
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 8 Covert Timing Channels OPtimal Capacity (OPC) send information as fast as possible E(X) is small (1,000s of packets/second) Fixed-average Packet Rate (FPR) send information as fast as possible with a fixed-average packet rate E(X) is fixed (a few packets/second)
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 9 Outline Background Covert Timing Channels Model-Based Framework Experimental Evaluation Capacity Detection Resistance Conclusion
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 10 Model-Based Framework The Framework filters and analyzes legitimate traffic encodes and transmits covert traffic
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 11 Components Filter filters input for the specified type of traffic (e.g., outgoing HTTP) outputs legitimate IPDs
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 12 Components Analyzer fits the legitimate IPDs to several models using MLE (blocks of 100 IPDs) selects the model with the lowest RMSE
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 13 Components Encoder uses the IDF of the model generates covert IPDs that mimic the legitimate traffic
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 14 Encoding / Decoding 1. Continuize 2. Encode 3. Decode 4. Discretize
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 15 Components Transmitter sends out packets with covert IPDs Receiver and Decoder receive packets and decode message
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 16 Model-Based Framework Implementation Details components run in user space filter, encoder, transmitter written in C; plus inline assembly for RDTSC analyzer written in MATLAB
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 17 Outline Background Covert Timing Channels Model-Based Framework Experimental Evaluation Capacity Detection Resistance Conclusion
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 18 Experimental Evaluation Test Scenarios LAN, WAN East-to-East, WAN East-to-West LANWAN-EEWAN-EW distance0.3 mi525 mi2660 mi RTT1.7ms59.6ms87.2ms IPDV2.5e e-032.1e-04 hops31813 IPDV – inter-packet delay variation
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 19 Test Setup MB-HTTP Weibull – avg. λ = , avg. k = E(X) is (~3 packets/second) OPC E(X) is 7.31e-3 to 7.87e-5 (1,515 to 12,777 packets/second) FPR Exponential – λ = E(X) is (~3 packets/second)
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 20 Theoretical Capacity channel LANWAN-EEWAN-EW CPPCPSCPPCPSCPPCPS MB-HTTP OPC 0.506, FPR CPP – capacity/packet, CPS = capacity/second LAN, WAN East-East, WAN East-West OPC has highest capacity
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 21 Theoretical Capacity channel LANWAN-EEWAN-EW CPPCPSCPPCPSCPPCPS MB-HTTP OPC 0.506, FPR CPP – capacity/packet, CPS = capacity/second LAN, WAN East-East, WAN East-West MB-HTTP and FPR are close
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 22 Empirical Capacity WAN East-East MB-HTTP versus FPR capacity and bit error degrade quickly
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 23 Empirical Capacity WAN East-West MB-HTTP versus FPR capacity and bit error degrade slowly
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 24 Empirical Capacity channel LANWAN-EEWAN-EW CPPCPSCPPCPSCPPCPS MB-HTTP OPC , ,512 FPR CPP – capacity/packet, CPS = capacity/second LAN, WAN East-East, WAN East-West OPC again has the highest capacity
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 25 Empirical Capacity channel LANWAN-EEWAN-EW CPPCPSCPPCPSCPPCPS MB-HTTP OPC , ,512 FPR CPP – capacity/packet, CPS = capacity/second LAN, WAN East-East, WAN East-West MB-HTTP and FPR are still close
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 26 Tests of Shape: Kolmogorov-Smirnov test – where s 1 and s 2 are distribution functions Tests of Regularity: The regularity test (Cabuk 2004) – 26 Detection Resistance
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 27 KSTEST LEGIT-HTTPMB-HTTP FPROPC sample size meanstddevm.s.d.m.s.dm.s.d 100x2, x10, x50, x250, KSTEST scores high mean and low s.d. for FPR and OPC
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 28 KSTEST LEGIT-HTTPMB-HTTP FPROPC sample size meanstddevm.s.d.m.s.dm.s.d 100x2, x10, x50, x250, KSTEST scores similar mean and s.d. for LEGIT and MB-HTTP
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 29 KSTEST KSTEST distribution similar distributions for LEGIT-HTTP and MB- HTTP scores
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 30 KSTEST KSTEST distribution LEGIT-HTTP and MB-HTTP overlap even with 250,000 packets
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 31 KSTEST LEGIT-HTTPMB-HTTP FPROPC sample sizeFPTP 100x2, x10, x50, x250, KSTEST detection rates FPR and OPC are detected easily
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 32 KSTEST LEGIT-HTTPMB-HTTP FPROPC sample sizeFPTP 100x2, x10, x50, x250, KSTEST detection rates FP equals TP for LEGIT and MB-HTTP
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 33 regularity LEGIT-HTTPMB-HTTP FPROPC sample sizemean 100x2,000 w= x2,000 w= regularity scores similar mean for LEGIT and MB-HTTP
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 34 regularity LEGIT-HTTPMB-HTTP FPROPC sample sizeFPTP 100x2,000 w= x2,000 w= regularity detection rates MB-HTTP is not detected at all
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 35 regularity LEGIT-HTTPMB-HTTP FPROPC sample sizeFPTP 100x2,000 w= x2,000 w= regularity detection rates again FPR and OPC are detected easily
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 36 Outline Background Covert Timing Channels Model-Based Framework Experimental Evaluation Capacity Detection Resistance Conclusion
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 37 Conclusion Model-Based Covert Timing Channels can be built automatically effective even in coast-to-coast scenario capacity is very close to FPR much stronger detection resistance than FPR and OPC
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 38 Conclusion (cont.) Future Work investigate detection methods for model- based covert timing channels explore other more advanced covert timing channel designs (e.g., non-parametric models)
RAID 2008 Model-Based Covert Timing Channels: Automated Modeling and Evasion 39 Questions? Thank You!