Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

1. Red Teaming Approaches, Rationales, Engagement Risks and Methodologies 2012 Information Systems Security Organization (ISSA) Information Security Forum Dave and Buster's 180 E Waterfront Dr Homestead, Pennsylvania Presenters: Mark YanalitisCISSP, IT Infrastructure Architect, PNC Bank Bill Johnson CPP, Director of Corporate Security and Employee Safety, Highmark Inc.

2. Goals for today • Define Red Teaming and its’ rationale • Discuss differences between commercial and full-spectrum Red Teaming • Discuss differences between commercial and full-spectrum methodologies • Examine common engagement risks • Application of Red Teaming methods • A companion document exists as supplemental reading resource for this presentation Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

3. Red Teaming Definitions “An array of activity where the overall goal is to understand the adversaries perspective in order to identify one's own vulnerabilities and challenge one 's own assumptions.”1 “Authorized, adversary-based assessment for defensive purposes.“2 “Review of control design and threat-based penetration testing to simulate actual attacks.”3 Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

4. Commercial v. Full Spectrum Lets explore the motivating factors for commercial and full-spectrum red teams Commercial engagements Full-Spectrum engagements Capabilities-based or hybrid modeling Simulations Goal is understanding Risk analysis driven Human in the loop Expand the knowable by parsing the unknown • Threat-based modeling • Compliance mandates • IT Audit adjunct testing • Goal is quick penetration • Cost and time driven • Automation dependency • Survey the known Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

5. Methodologies Commercial Full Spectrum Client Engagement Master Service Agreement Statement of Work Rules of Engagement Bond of Indemnity Identify Scope Reconnaissance Source: Sandia National Laboratories IDART, 2011 Targeting Reload Scan & Attack Compromise Report Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

6. The Intelligence Process 4Schudel, G. and Wood, B. (RAND, SANDIA & GTE: 2000) Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

7. Intelligence Cycle A full-spectrum red team will focus upon likely adversarial courses of action as well as current capabilities. Internally to the team, a need exists to have common doctrinal understanding of resource identification, intelligence collection, collection management, training, and leadership. Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

8. Intel Fusion Approach A red team collects and produces intelligence at variable rates and differing fidelities. The red team leader must be prepared for these eventualities. 5 Adapted from Steele, Robert, D. (New Craft of Intelligence: 2002) Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

9. Bias in the Intel process Sources of cognitive error Organizational and environmental bias indicators The assignment is not taken seriously The team or sponsor becomes too removed from the decision-making process A lack of interaction with the blue team Insufficient access to the details of the target Loss of team confidences The team fails to capture the details of the adversary, and instead mirrors itself The red team does offers no challenge to the blue team Thin top cover: the lack of a robust channel to act on findings in a timely manner, or consider findings with any seriousness. Applied post-event after many bodies already have been thrown at the problem The wrong team targeting the wrong problem (Threat-based team vs. a capabilities based problem) A lack of clarity on the urgency of issues at hand The red team approach is a one-time activity Sources of cognitive error can be found in individual minds, the collective agreement of the team, team composition, and in the quality of support given to the effort. Each individual team member carries both a cognitive bias as the known outsider pre-judged by their own past experiences, as well as the bias of their culture. 6 Defense Science Board. Task Force Report on The Role and Status of Red Teaming Activities: (DoD: 2003) Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

10. Role play the Adversary Lets explore the difference between a threat-based adversary model and capabilities-based adversary model. Threat– A threat represents a known quantity, a known effect singular in origin, essentially a Pathogen-Antigen model. A threat is an X-Y direct, or inverse relationship. Capability – Actors (or a confederation of multiple actors) capable of achieving a singular goal either due to access to resources, or some form of institutional support. The force multiplier effects of capability-based actors behave like an algebraic expression where leading factors have orders of magnitude, possibly even having orders of operation. Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

11. Adversarial Modeling The Universe of Actors and Actions In the full-spectrum Red Team context, the sponsor may need more than one type of red team to realistically model the capability. In the commercial world, modeling capability-based actors is the exception, not the norm. Red Team A Adversary B Red Team B Adversary A Adversary C Source: Sandia National Laboratories IDART, 2011 Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

12. Adversarial Prototyping ‥ 7Op. Cit.Schudel, G. and Wood, B. : 2000 Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

13. Threat Profiling 8 Duggan, et. al. SANDIA, 2007 Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

14. Attack Trees 9Schneier, Bruce. Modeling Security Threats (Dr. Dobb’s Journal: 1999) Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

15. Smart Grid Attack Diagram Corporate Network The Internet DDR PSTN DID 1-(724)-got-powr 128K CSU/DSU Link over carrier 10 Penn State University SIIS Laboratory. (Network and Security Research Center: 2010) Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

16. Smart Grid Attack Tree 11Ibid. Penn State University SIIS Laboratory: 2010 Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

17. When to Use Red Teaming 12 Atkins, William. Read Teaming – It's Good to be Bad. (Missouri S&T ACM SIG in Security: 2010) Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

18. Ok, you try it now. Target: A reciprocating high-speed gas compressor Source: BPI Compression 2011 Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

19. Case Study Client Situation A new foreign multi-national corporation (OCONUS) enters Southwestern PA as a result of a purchase and repackage of a number of leases sold by Chesapeake Energy. Almost all of these functional wells remain capped, and lack the infrastructure to get the gas to market. The corporations‘ local CONUS office has orders to open several fields and secure the infrastructure needed to handle gas compression and transfer. The local office issues an RFP, and engages a low-bid contractor solution. The preferred corporate contracted insurer does not feel the site protection solution adequately protects their underwriting investment of leased equipment. The insurer will not offer favorable rates until the site security concerns resolve to their satisfaction. Home office Corporate Security comes into the conversation late, and recommends that the site protection solution for the compressor be turned over to a Red Team/Blue Team for evaluation. Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

20. The red team simulation revealed complete destruction of the compressor and block house site plan in 10 minutes. What was the solution? Where are the vulnerabilities? why did the solution work? Flood light and CCTV camera ½ ton pickup A cinderblock Fender Jack Stump Remover Mentos Duct Tape Magnesium Ribbon 2 black Super Fan Suits Five 2L bottles of Cola One 2L bottle of Clorox 1 Large Sling Shot Estimated 3 weeks of site observation and rehearsal It frequently rains here No buried fence Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

21. Case Study Red Team Solution The full red team observed and collected intelligence on the target site for 3 weeks developing a number of information collections and applying information analysis techniques. They chose their best scenario. During the preparation phase, the pair melted down Stump Remover, which contains Potassium Nitrate, the main ingredient in smoke bombs. By adding in some refined sugar, the two made 2 baseball size smoke bombs on a kitchen stove. Magnesium ribbon, bought at an online camping store with a stolen credit card, served as the wicks.  The Cola was cut with 20% Clorox and the tops were taped over with Mentos in the bottle necks then loosely capped. The two planned to tape the bottle bombs into the seats and floors of the truck on site, and tie in place the steering wheel. Two red team members drove a rental car and a stolen pickup to the nearby target site. The red team members changed into black superfan suits (to prevent any hair, shoe, or clothing fibers from falling into the truck). The pair drove lights-out, a sanitized 3/4 ton stolen pickup truck (interior scrub-down, second stolen plate, and tire replaced with junk bald tires) to within 30 yards of the block house on a rainy night.  On site, the truck rear was jacked up.  While one threw/shot the lit smoke bombs at the base of the fence near the camera, the other placed the cinder block on the accelerator and turned on the ignition, and put the truck in drive.  When enough smoke obscured the camera, the jack was kicked out and the truck sped forward ramming the fence and crashing into the block house at high speed.  The Mento/Cola/Clorox bombs burst inside the cabin of the truck spoiling the interior environment defeating most physical forensic analysis.  The truck destroyed the compressor, and brought most of the block house down on the target.  The two fled on foot to the get-away car, and then later burned their suits in to black plastic lumps and discarded them. Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

22. Case Study: Blue Team Counter-Measures Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

23. Case Study: Blue Team Counter-Measures Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

24. Thank you for listening, are there any questions or comments? Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

25. Use your powers for the greater good, not evil. Fight the Good Fight Red Teaming Approaches, Rationales, Engagement Risks and Methodologies

26. References • McGannon, Michael. Developing Red Team Tactics, Techniques, and Procedures. (Red Team Journal: APR 2004). Internet. Found at http://redteamjournal.com/ • Sandia IDART Methodology. http://idart.sandia.gov/methodology/index.html • Price Waterhouse Coopers.Is your critical infrastructure safe? (PWC LLP:2010) Internet. Found athttp://www.pwc.com/en_US/us/industry/utilities/assets/cyber-attacks.pdf 5. • Schudel, G. and Wood, B. Modeling the Behavior of a Cyber Terrorist. (RAND National Security Research Division proceeding of workshop. Appendix C: Santa Monica, California: 2000) 49-59. Internet. Found at http://www.csl.sri.com/users/bjwood/cyber_terrorist_model_v4a.pdf. • Steele, Robert, D. The New Craft of Intelligence: Achieving Asymmetric Advantage in the Face of Nontraditional Threats. (U.S. Army War College Strategic Studies Institute: 2002). 34-36. • Department of Defense Science Board. Task Force Report on The Role and Status of Red Teaming Activities (Office of the Under Secretary of Defense For Acquisition, Technology, and Logistics. Washington, D.C. 20301-3140:2003). Internet. Found at www.au.af.mil/au/awc/awcgate/dod/dsb-redteam.pdf • Op. Cit. Schudel and Wood, 2000. • Duggan, David, P., Thomas, Sherry R., and Veitch, Cynthia K.K., and Woodward, Laura. Categorizing Threat - Building and Using a Generic Threat Matrix. (SANDIA National Laboratories, Albuquerque NM. REPORT SAND2007-5791: September 2007). Internet. Found at http://idart.sandia.gov/methodology/materials/Adversary_Modeling/SAND2007-5791.pdf • Schneier, Bruce. Modeling Security Threats - Attack Trees (Reprint Dr Dobb’s Journal: 1999. Counterpane Internet Security: 1999). Internet. Found at http://www.schneier.com/paper-attacktrees-ddj-ft.html • Penn State University SIIS Laboratory. Advanced Metering Infrastructure Security. (Penn State University Systems and Internet Infrastructure Security Laboratory, Computer Science and Engineering (CSE) Network and Security Research Center (NSRC): 2010). Internet. Found at http://siis.cse.psu.edu/smartgrid.html • Ibid. Penn State University, 2010 • Atkins, William. Read Teaming – It's Good to be Bad. (Missouri S&T ACM SIG in Security. SANDIA Critical Infrastructure Systems Department, NM, 10 FEB 2010). Internet. Found at. http://acm.device.mst.edu/security-files/2010-02-10-Red_Teaming.ppt ‡ See the supplemental paper that accompanies this presetation titled : Yanalitis, Mark. RED TEAMING APPROACH, RATIONALE, AND ENGAGEMENT RISKS (self-published: 2011). Red Teaming Approaches, Rationales, Engagement Risks and Methodologies