A production machine’s value and its controller’s security are separate questions. If a machine still meets production requirements but its controller cannot readily be patched, neither immediate replacement nor unrestricted connectivity should be the default. Start by separating the work the machine performs from the connections used to load programs, collect data and provide maintenance.
That distinction creates room for protection outside the controller. NIST’s guidance for small manufacturing environments gives a concrete example: where a device cannot support endpoint detection and response software, a firewall around it can serve as a compensating control. The protection sits at the boundary rather than inside the embedded device.
The aim is to preserve necessary production functions while reducing the ways vulnerable equipment can be reached. It is not to declare an old controller secure because it sits behind a firewall.
Make the machine’s connections visible
In an NC State Industry Extension Services webinar, the presenter describes legacy devices that were not intended for internet connectivity and may lack software updates or vulnerability patches. The discussion also challenges the assumption that an engineering colleague who manages computers necessarily owns policy enforcement, asset documentation and patching responsibility.
For a legacy machine review, make those responsibilities explicit. Ask production, maintenance, engineering, IT and the equipment supplier to describe the same installation. Build a connection record that distinguishes:
- The production equipment: its controller, production function and operating constraints.
- Supporting computers: any workstation used to prepare or transfer programs, collect machine data or run maintenance tools.
- Routine communications: which systems need to exchange information, in which direction and for what purpose.
- Maintenance access: how internal technicians and suppliers reach the equipment, including temporary connections.
- Ownership: who approves connectivity, maintains the boundary controls and authorizes equipment changes.
For example, if a CNC machine receives programs from a transfer computer and sends status information to a production application, document those exchanges separately. Do not turn “the machine needs the network” into permission for every office computer or supplier account to reach it.
Begin with existing drawings, configurations, maintenance records and discussions with the people who operate the equipment. Treat any proposed discovery tool or active scan as an equipment change requiring review, not as an automatic extension of office IT practice. The immediate objective is an agreed account of necessary connections, not a scan result obtained without operational approval.
Restrict communication outside the controller
Manufacturing network segmentation means separating equipment into controlled network areas and governing communication between them. For obsolete devices, NCSC’s secure connectivity guidance recommends isolation from the wider OT network, restrictions to operationally necessary connectivity, and maintained, security-hardened components between obsolete devices and external systems.
Apply that reasoning to the connection record. If a transfer computer must communicate with a controller, define that path specifically. Review whether the controller needs access to other machines, business systems or external networks rather than inheriting broad access from its current location.
The supporting computer deserves its own review. Identify its production role, other network connections and maintenance requirements. Keeping it within the review prevents the security boundary from being drawn around the controller while ignoring the system used to reach it.
NIST’s manufacturing segmentation guidance explains why indirect protection matters: embedded controllers, limited processing power and tight timing requirements can prevent the use of controls suitable for ordinary IT systems. A boundary control offers a different place to enforce restrictions, without assuming that security software can be installed on the machine.
Segmentation does not repair vulnerable software. It restricts access to that software. NCSC says obsolete products should be treated as untrusted and should not themselves implement security controls. It treats segmentation as temporary risk management while organizations assess sufficiency and establish a replacement timeline.
Give suppliers a controlled route in
Maintenance access should have a defined route, not become a standing exception to the machine’s network restrictions. NCSC recommends brokered vendor access through a secure gateway in a separate controlled network segment. The external party connects to an intermediary, which relays access to the OT asset. The intermediary must itself be actively updated and use modern authentication.
This separates the supplier’s entry point from the controller. It also places access enforcement on a maintained component rather than asking an obsolete controller to provide modern security capabilities.
Describe supplier access in operational terms: which technician needs access, to which equipment, for what task, and during which approved window. Include this in the plant’s identity and access management responsibilities so that ownership covers supplier access as well as employee accounts.
CISA’s joint guidance on adapting zero trust to OT recommends granting privileged access only when needed and for a limited duration, where this does not compromise safety or integrity. It identifies narrowly specified maintenance and support windows as a use of this approach, while warning that access must remain timely enough to avoid unsafe workarounds.
Agree the access arrangement with maintenance and the supplier before it is needed. Review whether the route reaches only the intended equipment, how access is enabled and ended, and who reviews activity. NCSC also recommends logging and monitoring interactions with obsolete devices.
Coordinate security changes with production
A security change is not successful if its deployment creates an unacceptable production or safety problem. NIST’s OT security guide explains that stopping and restarting equipment can affect production, and that familiar IT actions such as rebooting may conflict with OT availability and reliability requirements. It also calls for update testing by both the vendor and the end user.
Before changing a controller, supporting computer or communication rule, agree who will assess production effects, what will be tested, when the work can occur and how the previous configuration can be restored. Ask specifically what happens if a newly controlled connection becomes unavailable. NCSC’s connectivity guidance calls for assessing effects on core functions, safety, availability and new dependencies.
When a controller cannot be patched promptly, record the reason and the controls that restrict its exposure. CISA recommends testing patches outside the operational environment where feasible, requesting vendor testing information when that is impractical, and emphasizing isolation and compensating controls when prompt patching is not possible.
Do not extend that deferral automatically to the protective boundary. CISA calls for a more frequent patching cadence for edge devices and remote-access systems because of their role as security enforcement points.
Keep continued operation a deliberate decision
Retaining the machine should mean accepting a defined arrangement: necessary communications are documented, access is restricted, maintenance is coordinated and someone owns the remaining exposure. Revisit that arrangement when connections, suppliers, software support or production requirements change.
If a necessary connection cannot be controlled, or the components enforcing the boundary cannot be maintained, reconsider the connection, the support arrangement or the equipment replacement plan. Production usefulness alone should not settle the connectivity decision.
These preventive controls are separate from deciding what to do during a suspected compromise. A manufacturing ransomware tabletop can address those incident-time decisions without turning routine machine protection into an incident-response exercise.
If ownership of machine access and network boundaries remains unclear, explore Digiventi’s cybersecurity support to help frame the review around production needs and security responsibilities.