Plan the trust model and prove that certificate operations work

PKI Overview: Trust, Certificates and Lifecycle Responsibilities

A public key infrastructure needs more than a certificate authority.

  • 01Documented trust model and certificate use cases
  • 02Agreed issuance and integration workflows
  • 03Tested renewal and revocation scenarios
  • 04Operating instructions with owners and unresolved risks
11+Years of experience
120+Employees
540+Projects
ISO 27001certified

What does a PKI do, and where does its protection stop?

A public key infrastructure uses certificates, keys, policies and operating procedures to establish and check trust. A certificate binds identity information to a public key under an issuer's rules. Applications must still validate the certificate and make the correct authorisation decision. Possessing a certificate does not automatically grant access to every resource.

Private PKI establishes trust within the systems that are configured to trust its issuers. Publicly trusted certificates serve a different trust context. The design must match the actual use case, such as service authentication, device identity or document signing; those uses do not share every operational or regulatory requirement.

We start with the intended identities, applications, existing certificate authorities, key-storage arrangements, network constraints and operating roles. Access to a production private key is not a default prerequisite. The engagement separates architecture, implementation and optional ongoing operation. Supported technologies, service hours, incident response and any SLA must be confirmed in the agreement.

Issuance is not deployment: A successful renewal request does not prove that an application is serving the new certificate. Revocation behaviour also depends on the relying application and its configuration. Both need explicit acceptance checks.

4 service modules

Trust architecture, lifecycle tests and an operating handover

01

PKI Strategy and Architecture

You receive a documented trust design covering the selected identities, issuers, certificate profiles, key-protection requirements and operating responsibilities.

  • Distinguish private and publicly trusted certificate use cases
  • Record trust distribution, dependencies and recovery assumptions
  • Identify design decisions and risks that need an accountable owner
02

Implementation and Integration

The agreed issuance and application workflows are implemented in the selected environment and handed over with test results and remaining limitations.

  • Agree enrolment, identity checks and application integration
  • Test valid, invalid and untrusted certificate cases
  • Document configuration changes and a recovery or rollback path
03

Operations and Managed Services

The operating handover defines how certificates are discovered, monitored, renewed and revoked. Ongoing management is included only where expressly contracted.

  • Assign ownership to certificate records and consuming services
  • Verify deployment after renewal rather than stopping at issuance
  • Record incident, revocation, recovery and escalation responsibilities
04

Training and Operational Handover

Nominated operators work through the agreed runbooks and an example change or incident, exposing missing access or unresolved dependencies before handover.

  • Explain everyday tasks for the actual operator roles
  • Rehearse renewal and the selected failure scenario
  • Record open actions, document ownership and review triggers

Worked example: Prove renewal at the consuming endpoint

A fictional internal API uses a certificate issued by a private CA. The worksheet separates successful issuance from successful deployment and records what the relying client actually accepts. All scenarios are illustrative.

Illustrative certificate-lifecycle checks, not a complete PKI design or evidence from a client project.
Lifecycle stepEvidence to captureAcceptance checkAccountable role
Issue a replacement certificateCertificate profile, issuer and change recordIdentity and permitted use match the approved requestPKI operator and approving owner
Deploy to the internal APIEndpoint observation and deployed certificate identifierA representative client sees the replacement certificate and validates the intended trust pathApplication owner with the PKI operator
Test an invalid or revoked certificateClient configuration, test conditions and observed resultThe application behaves according to its agreed validation and revocation policy; gaps are recordedApplication security owner
Prepare the next renewalInventory owner, monitoring route and runbookAn alert reaches a responsible operator and the tested change procedure is availableService owner

5 phases

From trust requirements to a tested certificate lifecycle

Completion means that the agreed trust paths and workflows have test evidence, consuming systems have been checked and operating tasks have named owners. It does not mean every certificate in the organisation has been discovered or every application supports the same revocation mechanism. The agreed inventory boundary and untested dependencies remain visible.

  1. Define the use cases

    Identify identities, relying applications and the required trust boundary.

    Agreed scope and inventory assumptions.

  2. Design the trust model

    Resolve issuer, certificate-profile, key-protection and recovery decisions.

    Reviewed design with open decisions.

  3. Implement the selected workflows

    Configure issuance and application integration in the agreed environment.

    Configuration record and integration tests.

  4. Prove lifecycle behaviour

    Exercise renewal, deployment and the agreed invalid or revoked-certificate cases.

    Test evidence with application-specific limitations.

  5. Hand over operations

    Rehearse a runbook and confirm monitoring and escalation owners.

    Versioned instructions and outstanding actions.

Sarah Richter

Your contact

Sarah Richter

Head of Information Security, Cyber Security

10+ years of experience, CISA, CISM, Lead Auditor, DORA, NIS2, BCM, Cyber and Information Security

A PKI needs an operating model as well as certificate technology. Before choosing a platform, clarify who issues and renews certificates, how private keys are protected, how certificates are revoked after a key compromise and how dependent services recover. An inventory without tested ownership and lifecycle processes leaves important failure scenarios unresolved.
ADVISORI expert opinion

22 QUESTIONS, BRIEFLY ANSWERED

Frequently asked questions about PKI Overview

What is a Public Key Infrastructure (PKI) and what is it used for?

A PKI combines certificate authorities, certificates, key protection, policies and operating procedures. It supports use cases such as authenticating services or devices and verifying signatures. The application must validate the certificate and apply its own access rules. A PKI is a trust mechanism, not a guarantee that an application or business process is secure.

How do the Certificate Authority (CA), certificates, and keys work together?

A certificate authority issues a signed certificate that associates identity information with a public key under defined rules. The corresponding private key must be protected. A relying application checks the certificate and its trust path against its configuration and policy. Identity verification, issuance and application authorisation are separate responsibilities.

What are the greatest challenges in operating a PKI?

Common operational challenges include incomplete inventories, unclear ownership, certificate deployment after renewal, trust distribution, private-key protection and recovery. The engagement should identify the selected systems and test the actual workflow. A managed service needs explicit service boundaries and response responsibilities.

How is the security of private keys ensured in a PKI?

Protection depends on key use, storage, access control and the consequences of compromise. A design may use hardware-backed protection where justified, but a hardware module alone does not settle authorisation, backup or recovery. Required controls and operator responsibilities must be agreed for the selected keys and systems.

What role do Certificate Revocation Lists (CRL) and OCSP play in PKI?

Certificate Revocation Lists and the Online Certificate Status Protocol can communicate certificate status. Whether and how a relying application checks that information depends on its implementation and configuration. Test the intended policy, availability assumptions and failure behaviour rather than assuming that publishing a revocation immediately blocks every client.

What are the best practices for certificate lifecycle management?

Maintain an inventory with accountable owners, intended use, certificate and issuer details, renewal method, deployment location and evidence references. Monitor the actual lifecycle, rehearse renewal and verify the consuming endpoint afterwards. Different certificate use cases may require different policies; one validity-period rule should not be applied to every private and public certificate.

What do we receive from a PKI engagement?

Depending on the agreed scope, the handover includes a trust design, configuration and integration records, lifecycle test evidence, runbooks and an outstanding-action list. Architecture work, implementation and ongoing operations are separate commitments. Supported platforms, service hours and any SLA must be stated in the offer.

Can we use this overview as an Azure implementation guide?

The overview explains the trust and lifecycle decisions that apply across environments. An Azure implementation needs its own architecture, identities, integrations and operational checks. Use the linked Azure page to examine that distinct task rather than treating the two pages as interchangeable.

What practical benefits can an enterprise PKI provide?

A PKI can give selected systems a consistent way to establish certificate-based trust and manage certificate lifecycles. Useful outcomes include known issuers, accountable owners and repeatable issuance and renewal procedures. Those outcomes depend on the actual integrations and operating controls. A certificate does not by itself prevent phishing, authorise every application action or demonstrate compliance with an entire law or standard.

What must a cloud or hybrid PKI design resolve?

Identify where identities are checked, keys are generated and protected, certificates are issued and trust is distributed. Then confirm which applications can reach the required services and how renewal and recovery work. A representative client should validate the intended trust path and reject an untrusted case. Cloud hosting does not remove the need for clear issuer, application and operating responsibilities.

How does certificate authentication fit a Zero Trust design?

Certificates can help authenticate a service, device or user within an agreed trust model. The application still needs an access decision for the requested resource and action. Identity proof, device state, policy and authorisation have distinct roles. Installing a PKI alone does not establish a complete Zero Trust architecture. Test a valid identity that should still be refused access to a particular resource.

Which introduction risks should the design review expose?

Examples include an incomplete inventory, uncontrolled trust distribution, unclear renewal ownership, excessive issuer permissions and recovery procedures that have never been rehearsed. A design review should record assumptions and identify who must resolve them. Integration tests need invalid and failure cases as well as successful issuance. An unresolved dependency is an open action, not evidence that the whole environment is ready.

What changes for device fleets and industrial systems?

Devices may have limited storage, intermittent connectivity, long service lives or restricted update paths. The lifecycle design must therefore address enrolment, key protection, renewal, replacement and end of service for the actual devices. A signed update or certificate-based connection can support a particular security objective; it does not by itself prevent all tampering or unauthorised machine operations. Test representative devices and their operating constraints.

How are regulatory and assurance requirements determined?

Start with the actual service, certificate purpose, jurisdiction and applicable obligations. Internal workload authentication and a qualified trust service involve different roles and assurance questions. Record the identified requirements, evidence and accountable approver before selecting controls. Laws, technical profiles and management-system standards are not interchangeable. Neither a certificate nor a successful PKI installation is proof that every obligation has been met.

What can certificates contribute to email and document protection?

Suitable certificates and applications can support encryption or digital signatures. Encryption requires the intended key and recipient workflow; signature validation depends on trust, certificate status and the relevant application policy. Key loss, recovery and continued access to archived encrypted material must also be considered. A valid signature does not establish that a document's statements are true or that every legal form requirement has been satisfied.

Which certificate tasks are suitable for automation?

Repeatable tasks may include approved enrolment, issuance, inventory updates, renewal and deployment. The workflow still needs identity checks, bounded permissions, error handling and an accountable owner. Test what happens when issuance fails, deployment is rejected or the consuming service continues to use the old certificate. Automation is useful when its full operating path works; a successful API call alone is insufficient acceptance evidence.

How can PKI controls limit misuse by authorised insiders?

Separate sensitive duties where appropriate, constrain issuer and key access, and retain usable records of important operations. The design should consider whether an authorised operator could issue an unintended identity or misuse a signing key. Revocation and key-compromise procedures need tests in the relying applications. A hardware module or a published revocation does not automatically stop every form of misuse or invalidate every active session.

Which developments should influence a PKI roadmap?

Priorities should follow the installed environment: incomplete certificate inventories, changing application integrations, automated lifecycle needs and cryptographic dependencies. NIST's crypto-agility work is a useful reference when planning how algorithms and implementations can be changed. Assess actual compatibility, data lifetime and migration constraints before selecting a transition. Choose a change when its security objective, compatibility and operating responsibilities are clear.

How can certificate workflows support CI/CD safely?

A pipeline needs a clearly identified requesting workload and permissions limited to its approved certificate or signing use. Issuance, deployment and signing are different operations with different key-handling requirements. Secrets should not be copied into source code or exposed in logs. Test denied requests, failed deployment and key-access boundaries as well as the successful path. The pipeline and PKI owners must agree who maintains the integration.

What evidence is needed before migrating an existing PKI?

Inventory issuers, trust stores, certificate uses and relying applications before defining the transition. A staged test should cover old and new trust paths, renewal, revocation, rollback and recovery. Do not assume private keys can or should be exported: reissuance or new keys may be appropriate. Retirement of old trust needs explicit checks of dependent systems and retained verification needs, rather than relying only on the successful setup of the new CA.

How can certificates support remote access?

Certificate-based authentication can be part of an agreed VPN, network or device-access workflow. The design still needs issuance identity checks, endpoint management, application access policy and a process for lost or replaced devices. Test both a legitimate user and a device or identity that should be refused. A trusted certificate is one input to the access decision, not a blanket permission to the corporate environment.

What remains to protect after mutual TLS is enabled?

Mutual TLS can authenticate peers and protect the transport when configured correctly. Applications still need authorisation for each relevant resource and action, appropriate request validation and any required replay or duplicate-action controls. TLS 1.3 early data has specific replay considerations. A useful acceptance test therefore includes a trusted peer attempting a forbidden action; successful certificate validation alone should not make that action succeed.