Automotive Cybersecurity Part - 5: Risk Assessment Frameworks
๐ง Understanding Automotive Risk Assessments
โRisk comes from not knowing what youโre doing.โ Warren Buffett
This post explores two industry standard frameworks HEAVENS and EVITA that help cybersecurity professionals perform effective threat and risk assessments for Cyber Physical Vehicles (CPVs).
๐ What is Risk?
A risk assessment quantifies potential cybersecurity threats by evaluating
- Vulnerabilities
- Threats
- Likelihood of exploitation
- Impact on assets
- Effectiveness of existing security controls
Common formulas
Risk = Threat * VulnerabilityRisk = Threat * Vulnerability * Asset ValueRisk = ((Vulnerability * Threat) / Countermeasure) * Asset Value
โ๏ธ HEAVENS Framework
The HEAVENS model consists of three steps
- Determine Threat Level (TL)
- Determine Impact Level (IL)
- Determine Security Level (SL)
๐ Threat Level Parameters
| Parameter | Options and Values |
|---|---|
| Expertise | Layman (0), Proficient (1), Expert (2), Multiple Expert (3) |
| Knowledge of TOE | Public (0), Restricted (1), Sensitive (2), Critical (3) |
| Equipment | Standard (0), Specialized (1), Bespoke (2), Multiple Bespoke (3) |
| Window of Opportunity | Low (3), Medium (2), High (1), Critical (0) |
Final TL Value is the sum of the four parameters.
๐ฅ Impact Level Parameters
- Safety
- No Injury (0), Light/Moderate (10), Severe (100), Life Threatening (1000)
- Financial
- No Impact (0), Low (10), Medium (100), High (1000)
- Operational
- Ranges from No Impact (0) to Major Disruption (100)
- Privacy & Legislation
- No Impact (0), Low (1), Medium (10), High (100)
๐ Security Level (SL)
Final security levels are derived from TL and IL. A matrix is used to derive SL from the TL/IL combinations.
๐ก๏ธ EVITA Framework
EVITA is an EU funded initiative that evaluates risks across
- ๐ง Safety
- ๐ต๏ธ Privacy
- ๐ฐ Financial Losses
- โ๏ธ Operational Impact
๐ฏ Severity Classes (0โ4)
| Class | Safety | Privacy | Financial | Operational |
|---|---|---|---|---|
| 0 | No injuries | No data access | No loss | No impact |
| 4 | Fatalities | ID leakage | Multi vehicle losses | Multi vehicle failures |
๐งฎ Attack Potential Factors
| Factor | Levels |
|---|---|
| Elapsed Time | โค1 day (0) to >6 months (19) |
| Expertise | Layman (0), Expert (6), Multiple Experts (8) |
| System Knowledge | Public (0) to Critical (11) |
| Window of Opportunity | Unlimited (0) to None (โ) |
| Equipment | Standard (0) to Multiple Bespoke (9) |
๐ฒ Probability Mapping
| Score Range | Attack Probability |
|---|---|
| 0โ9 | Basic (5) |
| 10โ13 | Enhanced-Basic (4) |
| 14โ19 | Moderate (3) |
| 20โ24 | High (2) |
| โฅ25 | Beyond High (1) |
๐ง Key Takeaways
- HEAVENS gives a structured method for scoring risk via TL and IL.
- EVITA emphasizes attack effort and severity across various categories.
- Both frameworks are invaluable for assessing cybersecurity in connected vehicles.
You can use these frameworks as part of your automotive pentesting checklist or integrate them into CI/CD security gates for embedded automotive software.