Root of Trust


The root of trust model defines how users and different CAs can trust one another.

  • Each CA issues itself a certificate
    • called a root certificate
      • is self-signed
      • uses an RSA key size of 2,048 or 4,096 bits or the ECC equivalent
      • serves as the trust anchor
      • can be used to sign other certificates issued by the CA
    • Installing the CA’s root certificate means that hosts will automatically trust any certificates signed by that CA

Single CA

  • single root CA issues certificates directly to users and computers
  • often used on private networks
  • problem with this approach
    • the single CA server is very exposed
    • If compromised the whole PKI collapses

Third-party CAs

  • most third-party CAs operate a hierarchical model
    • root CA issues certificates to one or more intermediate CAs
    • intermediate CAs issue certificates to subjects (leaf or end entities)
  • advantage:
    • different intermediate CAs can be set up with certificate policies
      • enables users to perceive clearly what a particular certificate is designed for
    • Each leaf certificate can be traced to the root CA along the certification path
      • referred to as certificate chaining or a chain of trust

Self-signed Certificates

  • PKI can be difficult or expensive to manage
  • Any machine, web server, or program code can be deployed with a self-signed certificate
    • has been signed by the entity that issued it, rather than by a CA
  • e.g., web administrative interfaces of consumer routers are often only protected by a self-signed certificate
  • useful in development and test environments
  • OS or browser will mark self-signed certificates as untrusted
    • user can choose to override this
  • very difficult to validate
  • should not be used to protect critical hosts and applications