Fortinet NSE7_ESN_AR-7.6: Fortinet ZTNA Design Patterns
Fortinet Zero Trust Network Access (ZTNA) narrows connectivity around a specific application, user identity, device identity, and endpoint security posture instead of treating a network location as sufficient trust. In a FortiOS 7.6 environment, FortiGate can act as the application access gateway while FortiClient Enterprise Management Server (EMS) provides managed-device context and security posture tags. A good design makes access decisions understandable, enforceable, and maintainable as the workforce and applications change.
On this page
- Define the application boundary before the access policy
- Understand the FortiClient, EMS, and FortiGate trust relationship
- Use posture tags as one signal, not a complete identity system
- Protect certificate lifecycle and the application gateway
- Test positive, negative, and degraded cases
- Keep the design operable and auditable
- A design scenario: external contractor access to one application
The named NSE7_ESN_AR-7.6 article is most useful as a secure-network design discussion; Fortinet’s exam catalog has evolved, and readers should check the currently available certification and exam names rather than assume this historical exam label can still be scheduled. The architectural fundamentals below remain valuable for security engineers planning and reviewing a ZTNA deployment.
Define the application boundary before the access policy
List the application owner, URL or service name, protocol, backend servers, authentication model, expected users, and data sensitivity. A web application with SAML authentication presents different demands from a legacy TCP service or a system accessed only through an administrative client. Know whether the protocol can be fronted by an HTTPS application proxy or needs TCP forwarding, and identify application dependencies that may not be visible from the initial login screen.
Design least privilege around a meaningful business application, not a broad internal subnet. When multiple services are linked, document why they must be reached together and which teams own each dependency. A narrow access policy can become unreliable if DNS, identity services, certificate validation, or backend routing were never included in the architecture.
Understand the FortiClient, EMS, and FortiGate trust relationship
Managed FortiClient endpoints report device information and security posture to EMS. Tagging rules classify endpoints according to supported attributes, and FortiGate receives the corresponding ZTNA tags through its EMS integration. Depending on the deployment, device certificates can provide additional evidence about endpoint identity. This is not the same as proving the human user is authorized for every application.
Ensure that the EMS connector, trust certificates, endpoint registration, and policy synchronization remain healthy. A posture tag based on an outdated endpoint observation should not be assumed to represent current compliance indefinitely. Plan how the system behaves when EMS or its connection to FortiGate is unavailable and which diagnostic records show why access was allowed or denied.
Use posture tags as one signal, not a complete identity system
Security posture may consider managed status, operating-system information, protection software, vulnerability indicators, or other supported checks. Different organizations will require different conditions, but the rule should be explainable in business terms. A device that has lost critical protection may need restricted access even if its user credentials remain valid.
Keep user identity, device trust, and application authorization separate. A trusted managed laptop does not guarantee that every logged-in user is entitled to finance or production-management applications. Integrate the relevant identity provider and MFA controls, maintain group ownership, and explicitly map who may reach each application.
Protect certificate lifecycle and the application gateway
Device certificates and EMS connector trust are operational dependencies. Track issuing authorities, renewal, revocation, trust chains, and the effects of certificate changes on both managed endpoints and the gateway. Distinguish the EMS server’s management certificate from certificates used for device identity and from the application’s own TLS certificate.
FortiGate ZTNA application gateways need the correct hostname and server mappings, backend health, matching policy, supported authentication methods, and appropriate inspection. Design high availability and capacity before a large rollout. A gateway can be secure in its access decisions but still fail users if certificate renewal, backend DNS, or failover behavior is not tested.
Test positive, negative, and degraded cases
A useful acceptance test includes an authorized user on a compliant managed device, an unauthorized user on the same device, a compliant user on an unmanaged device, a device that loses a required tag, an expired certificate, an unreachable backend, and a temporarily unavailable identity dependency. The test should prove both correct access and correct denial without assuming that a success screen alone establishes enforcement quality.
Capture the request’s application identity, user or device context, matching tags and policy, gateway decision, and relevant backend result. Make sure administrators can distinguish authentication failure from authorization denial and from ordinary network unavailability; otherwise support teams may weaken policy in an attempt to restore service.
Keep the design operable and auditable
For each protected application, maintain a compact map of its owner, ZTNA server, policies, trusted identity groups, posture conditions, certificate dependencies, backend destinations, and change process. Review it after application migrations, user-role changes, and device-management updates. Policies and gateways for retired applications should be deliberately decommissioned so they do not become unexplained access paths.
The broader network security platform problem is the same across vendors: identity-aware access must remain consistent with routing, logging, certificate management, and operational recovery. The strongest Fortinet ZTNA design is one in which a security reviewer can explain exactly what grants access, what revokes it, and what evidence proves the rule works.
A design scenario: external contractor access to one application
Suppose a contractor needs access to a project-management application hosted behind FortiGate but should have no general connectivity to the company’s private network. The owner defines the authorized user group, the exact application hostname and backend service, the expected device-management requirements, and the duration of the contractor’s access. FortiClient EMS can provide managed-device posture evidence, while the identity provider and application roles remain responsible for identifying the user and authorizing the task.
The acceptance test includes more than one successful login. The same contractor should be denied access from an unmanaged device if policy requires management; a compliant device belonging to a different employee should not inherit the contractor’s entitlement; an expired certificate should fail safely; and the gateway should record a reason for a policy denial. The team should also confirm that a backend outage is reported as an availability problem, not disguised as a security rejection.
During the project, support teams need a documented route for escalating a failed tag synchronization or certificate trust issue without disabling all posture checks. When the engagement ends, deprovision the contractor’s group membership, revoke unneeded access, and review whether the ZTNA server remains necessary. The narrow design is valuable precisely because the organization can state what access exists and remove it cleanly.