architecting-security

Zero trust design starts with identity-first architecture: enforce MFA and SSO, apply least privilege with just-in-time access, and micro-segment networks by default. Every access request is authenticated, authorized, and continuously validated through a policy engine.

STRIDE threat modeling starts by building data flow diagrams of the system, then applying the six categories (Spoofing, Tampering, Repudiation, Information Disclosure, Denial of Service, Elevation of Privilege) to each component. Document threats, score them with DREAD, and design mitigations.

NIST CSF is a risk-based, flexible framework with 6 functions suited for governance and compliance mapping, while CIS Controls provide prescriptive, prioritized safeguards across 3 implementation groups. Use CIS for a tactical baseline and NIST CSF for program-level risk management.

For brownfield systems, use a hybrid approach: keep existing perimeter controls like firewalls and VPNs while layering zero trust controls progressively. Segment critical assets first and modernize identity and access management before full migration.

Generate an SBOM in CycloneDX or SPDX format automatically in your CI/CD pipeline, then scan dependencies continuously with tools like Trivy or Grype. For SLSA, start at Level 1 provenance and progress to hosted builds on trusted platforms like GitHub Actions.

Defense in depth alone is insufficient for cloud-native and remote-first environments where the network perimeter no longer defines trust. These scenarios require zero trust overlays with continuous verification, device posture checks, and identity-based access instead of network-based trust.