Power distribution, with purpose.
A switchgear cybersecurity boundary is a design decision about which signals may cross between electrical protection and control equipment, operational networks, and external systems. It is not simply a firewall placed around a control panel. The boundary must account for the actual source of each command or indication, what the equipment does when communications fail, and who is authorized to change settings. For a new switchboard or switchgear package, resolve those questions alongside the one-line diagram and control schematics so the physical and digital designs agree.
Start with the sources of control and status
Identify every source that can affect operation: local operator controls, protective relays, metering devices, PLCs, remote terminal units, building-management systems, and supervisory control systems. For each point, document whether it is an input, output, or command; whether it is hardwired or communicated; and which device is authoritative. A displayed breaker position, for example, may come from auxiliary contacts wired to a controller or from a network message. Those sources are not automatically equivalent, particularly when a display informs an operating decision.
The source-state schedule should describe normal, abnormal, and unavailable conditions. Specify how open, closed, tripped, and indeterminate states are represented, and distinguish a valid state from a stale value or lost connection. Identify whether a remote command is permitted at all, what local/remote selection means, and whether an operator must confirm an action locally. These details let the design team assess the consequence of a bad, delayed, or missing signal instead of assuming that all data are trustworthy.
Separate protective action from network supervision
Protection and supervisory control have different roles. A protective relay may detect a fault and initiate a trip through a defined control circuit, while a network connection reports event data or permits an approved supervisory command. The project should state which protective functions remain autonomous and which, if any, depend on communication. Do not let a general statement such as “network-controlled” obscure whether a communication outage could block a trip, prevent an interlock, or merely remove remote visibility.
Hardwired logic can provide a clear boundary for selected permissives, trip paths, and interlocks, but wiring alone does not establish cybersecurity or make a function fail-safe. A hardwired input may still be affected by incorrect design, maintenance, or component failure. Conversely, a communicated indication may be suitable for monitoring when its loss is annunciated and does not alter protection. Qualified protection and controls professionals should determine the required logic, separation, and failure response for the actual system and applicable requirements.

Define communication interfaces and permitted traffic
For each interface, record the connected device, protocol, physical medium, direction of data flow, and intended use. State whether the connection is for monitoring, settings access, time synchronization, engineering configuration, or control. Identify network ownership and the boundary between the switchgear controls and a customer or third-party network. Protocol compatibility by itself does not establish a secure interface, and a gateway should not be treated as an adequate boundary unless its functions, configuration, and management responsibilities are defined.
Specify how access is constrained and governed without assuming that a particular security feature exists in every relay or controller. Confirm supported authentication, account management, logging, configuration controls, and update methods for the actual selected components. Where remote access is unnecessary, omitting the interface can reduce exposure and maintenance burden. Where it is required, the owner’s IT/OT security team should define approved connection paths and practices. Device capabilities, component availability, and project requirements must be validated before they are promised in a specification.
Decide what happens when a boundary fails
Failure behavior should be stated for loss of network communication, gateway or controller failure, invalid data, power loss to control devices, and a failed or stuck output as applicable. For each case, identify what remains available locally, which alarms are produced, and whether protection continues independently. A useful design decision is whether a lost remote command path leaves the equipment in its present state, disables remote operation while retaining local operation, or invokes another defined response. The appropriate choice depends on the equipment function and operating risk.
Consider security events as well as ordinary outages. An unauthorized or malformed command, an account lockout, or a compromised engineering workstation may require isolation of an interface, but isolation must not unintentionally disable required protection or essential local control. Avoid vague requirements such as “fail secure” without specifying the intended state and consequences. Final settings, work practices, and interpretation of codes or standards belong to qualified professionals familiar with the equipment and site.
Compare architectures and document the decision
A project may choose a largely standalone control design, a monitored connection with no remote commands, or a controlled supervisory interface that permits defined commands. Standalone designs limit external exposure but can reduce centralized visibility and require local access for some tasks. Monitoring-only interfaces improve awareness while limiting command paths, yet still need rules for data quality and access. Supervisory control can support operations but adds dependencies, configuration responsibilities, and consequences to assess. The right choice follows the operating and security requirements, not a default preference for more connectivity.
Make the decision auditable in the one-line diagram, elementary and interconnection schematics, communications architecture, point list, control narrative, and cause-and-effect or sequence documentation as applicable. Keep an interface register and identify responsible parties for account administration, settings approval, change control, backups, and incident response. For a purchaser requesting a new custom assembly, provide the system one-line, nominal voltage and equipment rating requirements, available fault-current or short-circuit information, control sequence, point list, protocol and network requirements, and applicable owner standards. Ratings and standards compliance must be confirmed for the specific order.
Use the boundary as a scoped design, not a guarantee
Consider a hypothetical facility that needs breaker status at a supervisory system but does not require remote breaker commands. The design team could evaluate hardwired status contacts or a monitored communications link, define an explicit lost-communications alarm, and keep protective action independent of the supervisory connection. This is not a universal recommendation: point accuracy, isolation, device capabilities, operator response, and the owner’s security rules may change the decision. The team should document why the selected source is authoritative and what users are expected to do when it becomes unavailable.
The cybersecurity boundary describes a particular design at a particular time; it does not certify an entire facility as secure or guarantee behavior under every failure or attack. New equipment can be built to specified control and interface requirements, and UL 891 may apply to low-voltage switchboard assemblies where appropriate, but it is not a cybersecurity certification or a general medium-voltage listing. Confirm applicable standards, component availability, ratings, and any U.S.-origin requirements for the order. For Dallas–Fort Worth projects or shipments elsewhere, the purchaser’s site network rules and qualified engineering review remain essential.
Specifications & planning
Review the details that matter for your application. Final specifications and availability should be confirmed for your project.
| Evaluation factor | Example question for engineering review |
|---|---|
| Command authority | Which device or operator may issue each permitted command? |
| Loss of communication | What local functions remain, and how is the loss annunciated? |
| Data source | Is each status point hardwired, network-derived, or independently verified? |
| Access governance | Who approves accounts, settings changes, and interface maintenance? |
