Building trust from the root up

PKI becomes easier to reason about when it is treated as trust-chain and lifecycle design, not as a sequence of certificate-generation commands.

  • OpenSSL
  • X.509
  • CRL
  • TLS
Site plan · PKI-009

Trust is the architecture

Generating a certificate is the mechanical part of PKI. The engineering work is deciding which authority is trusted, which authority is allowed to issue what, how long each layer lives, and how compromise is contained. A root CA is valuable because it is trusted broadly, which is also why it should be used rarely and protected aggressively.

An intermediate CA creates an operational boundary. The offline root can authorize an intermediate with constrained purpose and path length; the intermediate can handle routine issuance. Replacing a compromised intermediate is painful, but it does not require replacing every root trust anchor.

A chain must be both valid and available

Leaf certificates need appropriate subject alternative names, key usage, and extended key usage. The serving edge must present the leaf and required intermediate certificates in the right order. Clients need the root in their trust store, but the server generally should not send that root as though it were part of the delivered chain.

Validation errors become legible when each link is checked: issuer and subject relationships, signatures, constraints, validity windows, name matching, and the chain actually presented over TLS. A correct set of files on disk is not proof that the edge serves a correct chain.

Lifecycle is part of trust

Issuance without rotation, inventory, and revocation is incomplete. CRLs and OCSP have different availability and freshness trade-offs, but both force an important design question: how will relying parties learn that a credential should no longer be trusted? PKI is a distributed policy system; certificate generation is only its most visible artifact.

Further down

Keep digging.

  1. K8S-031

    Delivery as reconciliation

    GitOps is most useful when desired state, reconciliation order, and ownership boundaries are explicit enough to operate under failure.

    • FluxCD
    • Helm
    • Kustomize
    • GitOps
    2 min read
  2. EDGE-012

    Ingress is a routing problem

    Ingress failures become tractable when DNS, TLS, edge reachability, controller routing, service discovery, and workload health are tested as one path.

    • Gateway API
    • Istio
    • Traefik
    • DNS
    2 min read
  3. CI-024

    One pipeline primitive, many languages

    Reusable CI should standardize security and delivery behavior across languages while leaving language-specific build logic explicit.

    • CI/CD
    • Reusable workflows
    • Supply chain
    2 min read