ISO 21434 Lifecycle and TARA

Vault note, not reviewed against the source. Written in the knowledge vault on 2026-05-12 by models working with Stefan Coetzee and published as it stands, with private addresses, e-mail addresses and an employer name redacted. Check claims against the primary source before relying on them.

CSMS โ€” Cybersecurity Management System

Organizational level. Includes:

  • Cybersecurity policy and processes.
  • Roles and responsibilities.
  • Resources and competence.
  • Distributed cybersecurity activities across product lifecycle.
  • Continuous cybersecurity activities (post-deployment).
  • Audit and management review.

OEMs and suppliers maintain CSMS supporting product cybersecurity engineering.

Product lifecycle phases

Concept phase

  • Item definition โ€” what is being developed; assumptions; boundaries.
  • Cybersecurity goals โ€” high-level outcomes.
  • TARA โ€” threat analysis and risk assessment.
  • Cybersecurity concept โ€” initial cybersecurity strategy.

Product development phase

  • Cybersecurity requirements derived from cybersecurity goals.
  • Cybersecurity design โ€” architecture, components, interfaces.
  • Integration and verification โ€” assemble, test cybersecurity.
  • Validation โ€” validate cybersecurity goals achieved.
  • Release for production.

Post-development phases

  • Production โ€” secure manufacturing process.
  • Operations and maintenance โ€” incident management, vulnerability management, security updates.
  • Decommissioning โ€” secure decommissioning.

TARA โ€” Threat Analysis and Risk Assessment

Asset identification

Identify cybersecurity-relevant assets:

  • Components, interfaces.
  • Data (user data, vehicle data, control data).
  • Functions.

Damage scenarios

For each asset, what damage could occur:

  • Financial impact.
  • Operational impact.
  • Safety impact.
  • Privacy impact.

Threat scenarios

For each damage scenario, threat agents and methods:

  • Attacker profiles.
  • Attack paths.

Attack path analysis

How could threat scenarios manifest:

  • Entry points.
  • Vulnerabilities to exploit.
  • Steps to achieve threat scenario.

Risk determination

Combine attack feasibility + impact:

  • Likelihood.
  • Severity.
  • Risk level.

Risk treatment

For each significant risk:

  • Avoid, transfer, mitigate, accept.
  • Document treatment decisions.

Cybersecurity engineering activities

Throughout lifecycle:

  • Requirements engineering โ€” derived from cybersecurity goals.
  • Design โ€” secure by design principles.
  • Implementation โ€” secure coding, secure configuration.
  • Verification โ€” code review, static analysis, dynamic testing, penetration testing.
  • Validation โ€” does the product achieve cybersecurity goals?

Distributed cybersecurity activities

OEM + tier-N supplier coordination:

  • Cybersecurity requirements flow-down.
  • Supplier-side cybersecurity activities.
  • Cybersecurity interface agreement.
  • Joint TARA for shared interfaces.

Continuous cybersecurity activities (post-deployment)

  • Cybersecurity monitoring โ€” production fleet monitoring.
  • Cybersecurity incident response โ€” detect, contain, recover.
  • Vulnerability management โ€” identify, assess, address.
  • Security updates โ€” develop, distribute, install.

Aligns with UN R156 SUMS for software update management.

Cybersecurity interface agreements

Between OEMs and suppliers:

  • Define cybersecurity responsibilities.
  • Define cybersecurity requirements at interface.
  • Define information exchange (vulnerabilities, incidents).
  • Define joint activities (TARA, audits).

AI-feature considerations

For AI features (ADAS, autonomous driving):

  • TARA includes AI-specific threats (adversarial inputs, model poisoning, prompt injection for LLM-based features).
  • Cybersecurity goals address AI-related impact.
  • ATLAS-aware threat analysis.
  • ISO/PAS 21448 SOTIF complements for AI safety.

SRE and AI-agent fit notes

Vehicle-specific work; less directly relevant to typical SRE engagements unless serving automotive clients.

For automotive client engagements:

  • Vocabulary alignment with ISO 21434 phases.
  • TARA methodology familiarity.
  • AI feature threat modeling using ATLAS + automotive specifics.

Stefan-context implementation sketch

  • For automotive client engagements: ISO 21434 + TISAX + UN R155 combined framing.
  • Limited direct applicability for general SRE / AI work outside automotive sector.

See also