Thin clients and zero clients are often treated as dumb terminals that need no security attention, yet each one boots firmware, holds certificates, caches credentials and decides which virtual desktop the user reaches. A thin client fleet is a set of networked computers with an operating system, a management server with push rights over every device, and a long replacement cycle that leaves old firmware in service for years. Secure Boot bypasses, expiring root certificates and default management passwords all apply to these devices as much as to laptops. This guide covers the device side of a VDI or DaaS estate: device types and their attack surface, boot and firmware integrity, the operating system image and write filters, management servers and enrolment, network and certificate trust, authentication at the device, and lifecycle and disposal. It is for end-user computing teams, VDI architects and security engineers. The platform side (brokers, gateways, session hosts, redirection policy) is covered in the Virtual Desktop (VDI) and DaaS Security Guide.
Key takeaways
- A thin client is a managed endpoint: inventory it, patch its firmware and OS, and monitor it like any other device, even if it stores no user data.
- Boot integrity matters on the cheapest devices: CISA added an IGEL OS Secure Boot bypass (CVE-2025-47827) to the Known Exploited Vulnerabilities catalog in October 2025, and the 2011 Microsoft Secure Boot certificates expire during 2026.
- The management server is the crown jewel: it can push images, settings and certificates to every device, so protect its administrators, service accounts and enrolment process like a software distribution system.
- Trust the network by certificate, not location: use 802.1X with device certificates, strict server certificate validation for brokers, and approval before new devices receive configuration.
- Retire on the vendor's support date: firmware lines reach end of support while the hardware still works, and unsupported zero client firmware cannot be fixed.
Device types and what each exposes
The terms are used loosely. The useful distinction is how much general-purpose software runs on the device, because that sets both the attack surface and the controls available.
| Type | What runs locally | Typical management | Main security considerations |
|---|---|---|---|
| Zero client | Protocol-specific firmware (for example PCoIP), no general OS, no local apps | Vendor management console or on-device web interface | Firmware signing and support status, device administrator password, certificate trust store, management console exposure |
| Linux-based thin client OS | Hardened Linux with read-only system partitions, protocol clients, browser in some builds | Vendor management server | Secure Boot chain, signed images and apps, local root and setup access, SSH and remote shadowing settings |
| Windows IoT Enterprise thin client | Full Windows with a write filter and lockdown features | Vendor console, Intune or Configuration Manager | Windows patching, write filter servicing, local administrator accounts, endpoint protection |
| Converted PC | Thin client OS installed on existing laptops or desktops | Vendor management server | Mixed firmware quality, legacy BIOS settings, leftover local disks with old data |
| Browser or ChromeOS-style endpoint | Browser-based access to DaaS | Cloud device management | Browser policy, extension control, session token theft |
Whatever the type, the device handles the user's credentials or smart card, renders the desktop and redirects peripherals. An attacker with control of the device can capture keystrokes and screen content, steal session tokens and replay credentials (ATT&CK T1078 Valid Accounts), and use redirected USB storage for data theft (T1052.001 Exfiltration over USB).
Boot and firmware integrity
Two 2025 and 2026 developments show why the boot chain needs active management on thin client fleets.
- Secure Boot bypass in the KEV catalog: CVE-2025-47827 affects IGEL OS before version 11. The CVE record states that the igel-flash-driver module improperly verifies a cryptographic signature, so a crafted root filesystem can be mounted from an unverified SquashFS image. CISA added it to the KEV catalog on 14 October 2025. Fleets still running IGEL OS 10 should treat this as a reason to upgrade or retire. This is ATT&CK T1542 Pre-OS Boot territory, where persistence survives reimaging of the visible OS.
- Secure Boot certificate expiry: Microsoft's support article on Secure Boot certificate expiration lists the Microsoft Corporation KEK CA 2011 and Microsoft UEFI CA 2011 as expiring in June 2026 and the Windows Production PCA 2011 in October 2026, replaced by 2023 certificates. Microsoft states that devices not updated keep booting but stop receiving new boot-level protections and revocation updates. Linux thin client OSes boot through a shim signed under the UEFI CA, so vendors ship dual-signed shims; IGEL, for example, lists dual-signed builds in IGEL OS 11.11.150, 12.8.1 LTS and 12.9.0. See the Secure Boot Certificate Expiry 2026 guide.
- Zero client root certificates: HP's Tera2 PCoIP zero client firmware release notes state that the previous PCoIP Root CA certificate was set to expire in August 2026, and that firmware 25.10.1, 26.01.1 and 26.05.1 include a renewed certificate that must be installed manually. A firmware update alone is not the whole fix.
Practical controls: enable UEFI Secure Boot and set a BIOS or UEFI administrator password on every device (vendor management tools can set these centrally), disable boot from USB and network unless used for imaging, keep firmware and BIOS on a supported release, and only install vendor-signed images and add-ons. See the Firmware and Boot Security Guide and the Linux Secure Boot and Shim Guide.
Operating system image, write filters and local data
Thin client operating systems are designed to keep the local system in a known state. The security value depends on how that design is configured.
- Appliance-style lockdown: Dell's ThinOS 10 security brief describes an Appliance Mode default with no command shell and no ability to remotely view, alter or delete OS, application or configuration files, and a model where only Dell-tested and certified packages can be installed, with hash and signature validation of packages. Most Linux-based thin client OSes follow a similar pattern of read-only system partitions and signed add-ons. Keep developer or debug modes off in production.
- Windows write filters: Microsoft's Unified Write Filter (UWF) redirects writes to a virtual overlay that is cleared on reboot, and is intended for kiosks, thin clients and shared workspaces. It is available on Enterprise, Education and IoT Enterprise editions. Microsoft notes three points that matter for security: updates must be applied in UWF servicing mode, including antimalware signatures; with fast startup on, shutting down does not clear the overlay; and removable drives cannot be protected. Microsoft does not describe UWF as a security boundary, so it complements patching and endpoint protection rather than replacing them.
- Exclusions are where persistence hides: every file, folder or registry key on the UWF exclusion list survives reboot. Keep the list short, documented and reviewed.
- Local data encryption: IGEL OS 12 offers device encryption under Security > Device Encryption covering partitions that can hold user data, such as browser history and custom partitions, with TPM-backed options (TPM+PIN, TPM PCR, TPM PCR+PIN). ThinOS 10 encrypts system, package and cached configuration partitions with device-specific keys and keeps sign-on credentials in RAM. Turn on what the platform offers, especially where a local browser or cached profile exists.
- Local accounts: change default device administrator and setup passwords before deployment, manage them centrally, and make them unique where the platform allows. A shared setup password across thousands of devices is a fleet-wide credential. See the Kiosk and Shared Device Security Guide for shared-account patterns.
Management servers, enrolment and their identities
The management server (IGEL UMS, Dell Wyse Management Suite, HP Anyware management, Windows device management for IoT Enterprise) is the highest-value target in the estate. It can push an image, a configuration that points users at a fake broker, or a certificate that the fleet then trusts.
| Risk | Control |
|---|---|
| Rogue device enrolment receiving configuration and certificates | Hold new devices in a quarantine group until approved; Dell's ThinOS brief describes a default policy group that holds clients until an administrator approves them |
| Rogue certificate enrolment server | Pin the CA for SCEP; HP's 26.05.0 zero client firmware added a check of the SCEP server's CA certificate against the trust store to block rogue SCEP servers |
| Administrator account takeover | Directory-backed accounts with phishing-resistant MFA, role separation between image publishing and helpdesk, no shared admin logins |
| Management traffic interception | HTTPS with validated server certificates between devices and server; management servers on a dedicated network segment |
| Service account and API key misuse | Inventory directory bind accounts, database accounts and API keys the server uses; vault and rotate them |
| Malicious or broken image pushed fleet-wide | Signed images only, staged rollout rings, change approval for image and firmware assignments |
The server's directory bind account, database credentials and any API keys used by help desk tooling are non-human identities with fleet-wide reach. See the Service Account Security Guide.
Network, protocol and certificate trust
- Network admission: use 802.1X with EAP-TLS and a per-device certificate (issued through SCEP from the management server where supported) and place thin clients on a VLAN that reaches only brokers, gateways, the management server, DNS, NTP and the CA. Dell lists EAP-PEAP, EAP-LEAP, EAP-TLS and EAP-FAST support; prefer EAP-TLS. See the Network Access Control (NAC) Guide.
- Broker certificate validation: clients connect to the broker or gateway over TLS. ThinOS offers validation levels named High, Warning and Low; set the strictest level and distribute internal root certificates centrally, so that a user can never click through a warning to an adversary-in-the-middle host (T1557).
- Protocol encryption: RDP, HDX, Blast and PCoIP all support encryption. Disable legacy or unencrypted modes, and on zero clients choose the strongest security mode the session host supports.
- Remote shadowing and SSH: disable on-device VNC shadowing and SSH unless the help desk needs them; if needed, require user consent and management-server-only access, and log each session.
- Peripheral control at the device: allow or deny USB devices by vendor and product ID or class at the device as well as in the session policy, so a blocked class never reaches the session. See the Removable Media and USB Control Guide.
Authentication at the device
The device is where the user proves who they are, but the decision belongs at the broker or identity provider. Thin clients commonly support smart cards (PIV/CAC), FIDO2 security keys passed through to the session, and badge tap-and-go single sign-on for shared clinical or retail workstations. Enforce MFA at the gateway or identity provider so that a stolen password used from any device fails. The Change Healthcare breach began with a remote access login that had no MFA; the endpoint type did not matter. See the Smart Card and PIV/CAC Authentication Guide.
Choosing a device model
| Situation | Better fit | Trade-off |
|---|---|---|
| High-security, single protocol, no local apps needed | Zero client | Firmware lifecycle tied to one vendor and protocol; fewer controls such as local MFA options |
| Mixed protocols, local browser, smart cards, unified communications offload | Linux thin client OS | More software to patch; needs management server hardening |
| Windows-only peripherals or line-of-business apps that must run locally | Windows IoT Enterprise with UWF | Full Windows patching and endpoint protection burden |
| Extending life of existing PCs | Converted PC | Check firmware support, Secure Boot capability and wipe old disks first |
| Small organisation, cloud DaaS, no on-premises servers | Cloud-managed thin client OS or browser endpoint | Management depends on the vendor's cloud identity and MFA |
Implementation: a phased rollout
- Inventory: list every device with model, OS or firmware version, BIOS version, Secure Boot state, management group and location. Compare versions with each vendor's support lifecycle and the KEV catalog.
- Harden the management server first: MFA for administrators, role separation, HTTPS with valid certificates, quarantine for new devices and a backup of configuration.
- Build a baseline profile: Secure Boot on, BIOS password set, USB and network boot off, shadowing and SSH off, strict certificate validation, encryption on, USB allow-list, renewed root certificates installed.
- Pilot ring: apply the profile to IT devices, then a small group per site and peripheral type (scanners, signature pads, smart card readers). Test certificate validation against every broker URL.
- Roll out by site or ring, with the previous profile kept as the rollback assignment. Firmware updates go in rings too, with recovery images ready for devices that fail to boot.
- Replace what cannot be fixed: devices on unsupported firmware, without Secure Boot, or unable to take the 2023 certificates move to a replacement plan with a date. See the Operating System End of Life and Extended Support Guide.
- Dispose securely: reset to factory, remove from management, revoke device certificates and sanitise or destroy local flash. See the Media Sanitisation Guide.
Detection and measurement
Forward management server audit logs (administrator logons, image and profile assignments, firmware pushes, certificate deployments) and device logs where supported to the SIEM. Alert on new administrator accounts, profile changes outside change windows, devices enrolling from unexpected networks, devices reporting Secure Boot off, and 802.1X failures for known device certificates.
| Metric | Target |
|---|---|
| Devices on a vendor-supported firmware or OS release | 100% |
| Devices reporting Secure Boot enabled with current certificates | 100% of capable devices |
| Devices authenticating to the network with a device certificate | Above 95%, with exceptions listed |
| Management administrators with phishing-resistant MFA | 100% |
| Median days from firmware security release to fleet deployment | 30 days or fewer (NHI Mgmt Group recommendation) |
| Devices not seen by the management server in 30 days | Investigated and resolved |
Common failure patterns
- "Nothing to steal" thinking: devices are left out of patching and vulnerability scanning because they hold no data, while they hold credentials, certificates and the path to every desktop.
- Default setup passwords: the same device administrator password everywhere, printed in deployment guides.
- Open enrolment: any device that reaches the management server receives the production profile, including broker addresses and certificates.
- Certificate warnings left to users: low validation levels let a user accept a spoofed broker certificate.
- Long tails of old firmware: remote sites and spare stock keep devices on unsupported releases, including those affected by known Secure Boot bypasses or expiring root certificates.
- Write filter as a security boundary: UWF exclusions grow, servicing mode is skipped, and malware or stale signatures persist.
Framework mapping
| Control area | NIST SP 800-53 Rev. 5 | NIST CSF 2.0 | ISO/IEC 27001:2022 Annex A | CIS Controls v8.1 |
|---|---|---|---|---|
| Device inventory and supported versions | CM-8, SA-22 | ID.AM-01, PR.PS-02 | A.5.9, A.8.1 | Controls 1 and 2 |
| Boot, firmware and image integrity | SI-7 | PR.PS-02 | A.8.19 | Control 4 |
| Hardened configuration and lockdown | CM-7, CM-6 | PR.PS-01 | A.8.9 | Control 4 |
| Device authentication and network admission | IA-3 | PR.AA-03, PR.IR-01 | A.8.20 | Controls 12 and 13 |
| Protocol and management encryption | SC-8 | PR.DS-02 | A.8.24 | Control 3 |
| Remote support and shadowing | MA-4 | PR.PS-01 | A.8.9 | Control 4 |
Relevant ATT&CK techniques: T1542 Pre-OS Boot, T1542.003 Bootkit, T1200 Hardware Additions, T1052.001 Exfiltration over USB, T1078 Valid Accounts and T1557 Adversary-in-the-Middle.
Practitioner checklist
- Inventory every thin and zero client with firmware, BIOS and Secure Boot state, and compare against vendor support and the KEV catalog.
- Upgrade or retire IGEL OS releases before 11 and any device that cannot take the 2023 Secure Boot certificates.
- Install renewed protocol root certificates on zero clients where firmware notes require a manual step.
- Set BIOS or UEFI passwords, enable Secure Boot and disable USB and network boot.
- Protect the management server with MFA, role separation, HTTPS and an approval queue for new devices.
- Use 802.1X EAP-TLS with device certificates and a restricted VLAN.
- Set the strictest broker certificate validation and turn off shadowing and SSH unless needed.
- Enable device encryption and keep write filter exclusions short and reviewed.
- Revoke certificates, remove from management and sanitise flash at disposal.
Standards and references
- CISA: Known Exploited Vulnerabilities Catalog
- Microsoft: Windows Secure Boot certificate expiration and CA updates
- Microsoft Learn: Unified Write Filter (UWF)
- IGEL: Device Encryption in IGEL OS 12
- Dell Technologies: ThinOS 10 Security Brief
- HP Anyware: Tera2 PCoIP Zero Client Firmware Release Notes
- NIST: SP 800-46 Rev. 2 Guide to Enterprise Telework, Remote Access and BYOD Security
- NCSC: Device Security Guidance
Related NHI Mgmt Group resources: Virtual Desktop (VDI) and DaaS Security Guide · Endpoint Security Guide · Firmware and Boot Security Guide · Remote Desktop (RDP) Security Guide