cyber adversary characterization know thy enemy!
TRANSCRIPT
Cyber Adversary Characterization
Know thy enemy!
Introduction and Background
• Cyber Adversary Characterization workshop in 2002
• Research discussions continued via email
• Briefings to Blackhat and Defcon to introduce concept and obtain feedback
• Future workshops planned for October 2003
• Slides will be on both conference web sites
Why characterize?
• Theoretical: To gain understanding of and an ability to anticipate an adversary in order to build improved threat models.
• Practice: Improved profiling of attackers at post attack and forensic levels.
Point Scoring: Rating-the-Hacker
Toby Miller
Point Scoring: Why?
• No “standard” system to help rate the attacker
• No system to help with the threat level
• Help management in the decision making process
Point Scoring: The Categories
• Passive Fingerprinting
• Intelligence
• The Attack
• The Exploit
• Backdoors | Cover up
• Other
Example Score Metric
Linux 3
FreeBSD 4
OpenBSD 6
IRIX 4
Windows 3
Point Scoring: Past, Present, Future
• Originally posted on incidents.org
• Currently on rev2
• Soon to release rev 3
• www.ratingthehacker.net
Tool characterizations, Disclosure Patterns and
Technique scoring.
Tom Parker – Pentest Limited (UK)
The Hacker Pie
• Representative of characterization metrics which build the final characterization.
• Available elements dependant upon scenario.
• Does not rely solely upon IDS/attack signature data.
The Hacker Pie (continued)
• Pie reliant upon the results of multiple metrics which are, in many cases inter-related, strengthening the likelihood of an accurate characterization.
• Relationships between key metrics and key data enable accurate assumptions to be made regarding unobserved key information.
The Pie Explained
Metric One Metric ThreeMetric Two
Key Data Key Data Key Data Key Data Key Data
Characterization
Metric Four
021
2
Point Scoring Systems (Continued)
• Attempt to characterize an adversary based on attack information captured from the wild.
• Attempt to characterize adversary based upon “technique classification model”
• Attempt to characterize adversary based upon “tool classification model”
Tool classification model
• Availability of application
• Origins of application
• Ease of use– Requires in-depth knowledge of vulnerability to
execute?– Other mitigating factors
Example Exploit ClassificationWeb App Flaw Public PrivateProprietary Application Penetration Via SQL Injection 3 4Open Source Application Penetration Via SQL Injection 3 4Proprietary Application Penetration Via Arbitrary Script Injection 2 3Open Source Application Penetration Via Arbitrary Script Injection 2 3Proprietary Application Penetration Via OS command execution using SQL Injection (MS SQL)
3 5
Proprietary Application Penetration Via OS command execution using SQL Injection (other)
4 7Proprietary Application Penetration Via SQL Injection (MS SQL) 5 6Proprietary Application Penetration Via SQL Injection (other) 4 7
Disclosure Food Chain Characterization
• All tools have a story
• Often years before dissemination into public domain.
• Social demeanour often key to placing in disclosure disclosure chain.
• “Pyramid” metric.
Exploit Development
Vendor Coordination
Public Disclosure
Exploit Reverse Engineered / Vulnerability Research
Honey Pot Capture
Exploit Usage In Wild
Exploit TradingType title here
Vendor Patch Released
Public Disclosure
Vendor Coordination
Public Disclosure Vendor Fix Released
Further Research
Disclosure to Security Company
Information shared further throughout grey hat communities
Information shared with fellow researchers (Exploit Development)
Vulnerability Discovery
The Disclosure “Food Chain”
2 Approaches to Modeling the Cyber Adversary: Offender Profiling & Remote Assessment
Dr. Eric D. Shaw
Consulting & Clinical Psychology, [email protected]
Offender Profiling
• Roots in Law enforcement & intelligence community (criminal event or incident analysis)—intensive review of past offenders
• Insider Computer Crimes, 1998-present– 50 cases– 10 in-depth case studies from companies or gov’t. contractors
• Products– Typology of actors: motivation, psychological characteristics, actions– Critical pathway—process of interactions w/environment (personal and
professional) leading to attack– At-risk characteristics– Organizational vulnerabilities & Insights into prevention, deterrence,
detection, management
Offender Profiling Headlines
• The Termination Problem
• Actor subtypes—the Proprietor & Hacker
• The Tracking Problem
• Organizational Vulnerabilities
• Detection Issues
• Intervention Challenges
• Hacker Overview
Attacks: The Termination Problem • Simple termination of Disgruntled Insider is not
the answer—80% attack after termination (4 hours-2 months)
• 70% attack from remote locations vs. inside—termination did not impact access
• Attack types:– DOS to disrupt business
– Destruction & corruption of data
– Theft of Proprietary data
– Time bombs
– Extortion
– Attack on reputations
Attackers
• Hackers—40%: affiliated with and active in hacking community, brings hacking practices to worksite
• Proprietors—40%: defend system as belonging to them, resist efforts to dilute control
• Avengers—20%: attack impulsively in response to perceived injustice
Prevention: Screening & Selection
The Tracking Problem
• Screening & Selection Problems in 60% of cases—no or delayed background, nepotism, failure to detect risk factors
• 30% had prior felony convictions
• 30% had high-profile hacker activity
Organizational Issues
• 80% of cases occur during periods of high organizational stress or change at the highest to supervisory levels
• Lack of policies contributed to disgruntlement or facilitated attack in 60% of cases
• Lack of policy enforcement contributed to disgruntlement of facilitated attack in 70% of cases
Detection Problems
• 80% of attackers used operational security to protect attack planning or identity
• Time disgruntled to attack: 1-48 months with a mean of 11.3 months
• Time active problems (probation) to attack: 0-76 weeks with a mean of 26 weeks
Forget the “big bang” theory of the sudden, unforeseen attack
Intervention Problems
• Management intervention initially exacerbated problems in 80% of cases (ignore, placate or tolerate problems, negotiate then cut-off, terminate poorly)
• Problems with termination process in 80% of cases (esp. failure to terminate access)
• Multidisciplinary risk assessment prior to termination
Hardcore Hackers: Not Script Kiddies
AgeMean=25.5
Tech
Capability
Prior
Offenses50%
Acted with
Others75%
Status in Hacker Community
Oquendo 29 High Yes Yes High
Zezev 30 High No Yes Unknown
Carpenter 20 High Yes No Low
Demostenis 23 Low No Yes Low
Remote Assessment Using WarmTouch(patent pending)
Why Use WarmTouch Software to Detect Disgruntlement or Psych Change
on-line?
• Communication has moved on-line
• Loss of visual & auditory cues on-line
• Failure of other systems to detect violations: technical noise, supervisor & peer reporting
• Protects Privacy
• Provides Objectivity
VulnerableCITI
Minor Infraction
Moderate Infraction
MajorAct
Personal Stressors
Professional Stressors
Mounting Stress and Frustration
Person-Situation Interaction:Detect Psychological “Leakage”
“Software” Components• Psychological Profiling Algorithms
– Emphasis on measuring emotional state• Anger• Anxiety• Depression
– Changes in emotional state from baseline
• Psychological characteristics: decision-making and personal relations– Loner/team player– plans/reacts– Rigid/flexible– Sensitivity to environment
• Alert Phrases-key words– Threats– Victimization– Employment Problems
• Communication Characteristics– To, From, Time, Length, etc.
WarmTouch “Software” Overview
• WarmTouch origins in IC, 1986-present • Use of WarmTouch with Insider Communications
– Khanna at Bank– Threat Monitoring– Sting operations & negotiations– Suspect identification– Hanssen
• Other WarmTouch Applications
Case Example: Financial Proprietor
• Well paid systems administrator • Personality Traits-Proprietor
– Entitlement– Manipulative– Devaluing of others– Padded OT
• Context: Supervisor Change
Email from Boss
• Asked to train back-up
• “You seem to have developed a personal attachment to the System Servers. These servers and the entire system belong to this institution not to you…”
Email 1: April
• (Asked to train his back-up, subject refuses) “His experience was ZERO. He does not know ANYTHING about ...our reporting tools.”
• “Until you fire me or I quit, I have to take orders from you…Until he is a trained expert, I won’t give him access...If you order me to give him root access, then you have to permanently relieve me of my duties on that machine. I can’t be a garbage cleaner if someone screws up….I won’t compromise on that.”
Email 3: July
• “Whether or not you continue me here after next month (consulting, full-time, or part-time), you can always count on me for quick response to any questions, concerns, or production problems with the system. As always, you’ll always get the most cost-effective, and productive solution from me.”
Email 4: July
• “I would be honored to work until last week of August.”
• “As John may have told you, there are a lot of things which at times get “flaky” with the system front-end and back-end. Two week extension won’t be enough time for me to look into everything for such a critical and complex system.”
• “Thanks for all your trust in me.”
The Event
• On last day of work, subject disables the computer network’s two fileservers.
• Company executives implore subject to help them fix the problems, but he refuses.
• Independent consulting firm hired to investigate problems, discovers sabotage.
• Timing: deception to cover plotting.
WarmTouch Challenge
• Detect deterioration in relationship with supervisor
• Detect Deception
The April Email Profile
July Email Profile
• August
Detecting Deception
Covert vs. Overt Hostility in Email Prior to Attack
Three MonthsPrior
Two MonthsPrior
Two Weeks Prior
Attack
Overt Hostility
Covert Hostility
Zezev vs. Bloomberg: Managing his Psychological State
• Task: to lure him to London for the bust – must manage his anger and anxiety at delays and
manipulations– satisfy his dependency—need for $ & job
• Warmtouch help:– Objectively highlight and help manage
psychological states– Objectively measure success
Support to Sting Ops/Negotiations: Levels of Anger in Zezev’s emails to
BloombergIndicators of Anger (+)
0
50
100
150
200
250
300
350
400
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Evaluators -
Evaluators +
Feelings -
Feelings +
Direct Ref.
Negatives
Me
We
I
Zezev’s Use of “Me”passive/dependent mode
Me
0
0.5
1
1.5
2
2.5
3
3.5
1 3 5 7 9 11 13 15 17 19
Zezev’s Use of RetractorsAnxiety
Retractors
0
1
2
3
4
5
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Robert Hanssen
• 8 Communications with Soviet Handlers
• Between October 1985 & November 2000
• Challenge for Software:– Detect signs of emotional stress associated with
spying, disgruntlement and “affair” as documented in public records
Hansen: Anger over Time
Hansen: Changes over Time
0
5
10
15
20
Number of Words
10/1/1985 9/8/1987 6/8/2000
Date
Psycholinguistic Measures of Anger
NegativesMe
Hansen: Changes Over Time
05
101520253035404550
Number of Words
10/1/1985 11/8/1985 6/13/1988 6/8/2000
Date
Emotional Vulnerability
Adv Intensifiers
Direct Ref
Feelings
I
Hansen: Changes over Time
0
2
4
6
8
10
12
14
Number of Words
10/1/1985 11/8/1985 6/13/1988 6/8/2000
Date
Psycholinguistic Measures: Anxiety
Explainers
Retractors
Other WarmTouch Applications
• Communications Manager– Analyze state of relationship
– Assess characteristics of persons in relationship
– Help modify language to improve/modify relationship
– Track success/changes over time
• Media Monitoring– Attitude of Egyptian press toward U.S.
– Attitude of customers toward product or service
Internet Threat Actors
Marcus H. SachsDirector, Internet Storm Center
The SANS Institutehttp://isc.sans.org
• US national information networks have become more vulnerable—and therefore more attractive as a target
• Growing connectivity among secure and insecure networks creates new opportunities for unauthorized intrusions into sensitive or proprietary computer systems
• The complexity of computer networks is growing faster than the ability to understand and protect them
• The prospects for a cascade of failures across US infrastructures are largely unknown
The Cyber Threat to the United States
• Hacker/Script Kiddies/Hobbyist• Disgruntled Employee• Insider aiding others• Hacktivist• Industrial Espionage• Foreign Espionage• Terrorist• State Sponsored Attack
Cyber Threats to the Critical Infrastructure
Low High
High
Low
PotentialDamage
Probability of occurrence
2003
2004
2005
Source: 1997 DSB Summer Study
HackerCriminal
Espionage
Terrorist
State Sponsored
The Threat is Increasing
• Internet was not built to be secure• “Secure” (i.e., obscure) software being replaced by
commercial products in infrastructures• Software development focused on “Slick, Stable,
Simple” (not “Secure”)• System administrators lack training• Leaders rarely see computer security as part of the
“bottom line”• User awareness is low
Why are we so Vulnerable?
• The real threat to the Critical Infrastructure is not the hacker, but the structured state-sponsored organization
• However...– Sometimes it’s hard to tell the difference - both use the same tools– Growing sophistication and availability of tools increases concern– Must assume the worst until proven wrong
• So...– The government takes seriously all unauthorized activity– They will use all technical and law enforcement tools to respond ... and
deter– They will seek legal prosecution where appropriate
Why The Feds are Concerned About Hackers
http://www.whitehouse.gov/homeland/
New Homeland Security Strategies
National Strategy to Secure Cyberspace
• Nation fully dependent on cyberspace
• Range of threats: script kiddies to nation states
• Fix vulnerabilities, don’t orient on threats
• New vulnerabilities require constant vigilance
• Individual vs. national risk management
• Government alone cannot secure
cyberspace
• Enhance law enforcement’s capabilities for preemption, prevention, and prosecution
• Secure the mechanisms of the Internet including improving protocols and routing
• Foster trusted digital control systems/ supervisory control and data acquisition systems
• Reduce and remediate software vulnerabilities• Improve physical security of cyber
and telecommunications systems
Priority IIA National Cyberspace Security
Threat and Vulnerability Reduction Program
Inside the Internet Storm CenterData Collection
DShield Users
Analysis Dissemination
DShield.org
Typical Residential Cable Modem Log
Pop-up ads
(Spam)
FTP attempt
s
Pop-up ads
(Spam)
FTP attempt
s
Internet Storm Center Web Page
http://isc.sans.org
Port Report
2002 Top 20 ListTop Vulnerabilities to Windows Systems
W1 Internet Information Services (IIS) W2 Microsoft Data Access Components (MDAC) -- Remote Data Services W3 Microsoft SQL Server W4 NETBIOS -- Unprotected Windows Networking Shares W5 Anonymous Logon -- Null Sessions W6 LAN Manager Authentication -- Weak LM Hashing W7 General Windows Authentication -- Accounts with No Passwords or Weak Passwords W8 Internet Explorer W9 Remote Registry Access W10 Windows Scripting Host
Top Vulnerabilities to Unix Systems
U1 Remote Procedure Calls (RPC) U2 Apache Web Server U3 Secure Shell (SSH) U4 Simple Network Management Protocol (SNMP) U5 File Transfer Protocol (FTP) U6 R-Services -- Trust Relationships U7 Line Printer Daemon (LPD) U8 Sendmail U9 BIND/DNS U10 General Unix Authentication -- Accounts with No Passwords or Weak Passwords
www.sans.org/top20
