A useful starting point is the function that the plant needs to keep operating. For one line this may mean producing an approved product configuration, recording changes and resuming operation in a controlled manner after an interruption. That determines which systems, people and service providers need to be considered. The IEC 62443-3-2 method connects the system boundary to risk assessment and protection requirements. [7]
What boundary are we assessing?
The line may include the controller, operator panel and recipe management. It also depends on the transfer of production orders and on a maintenance service. For each relationship, we establish what is permitted in normal operation and what changes if a connection fails or data are incorrect. A safety-related function adds another question: Could a digital modification expose a person to hazardous movement?
What consequences are acceptable?
The plant may be able to complete an already loaded order for two hours without contact with the planning system. When a recipe is changed, however, the correct version is essential from the first second. Both requirements can be tested. A time limit without a link to a function, or a blanket demand for “high security”, is of little use during acceptance.
How do we limit the impact of a fault?
IEC 62443 groups equipment with shared protection requirements into zones. Intended communications between zones are considered as conduits. For one line, transferring production counts may have a different requirement from write-capable remote maintenance, which may need individual authorisation and time-limited approval. That distinction follows from the permitted activity and its possible consequences. A line on a network drawing says little by itself. [7]
How do protection measures work together?
A limited maintenance connection restricts what can be reached. Individual permissions define who may work there. Logging helps establish what was changed afterwards. An appropriately secured and tested project version enables a return to a known state. This is the principle of layered protection, often called “defence in depth”. Detection and recovery are as much part of a protection concept as preventing unauthorised actions.
What can the operator actually demonstrate?
A target level describes the protection that is sought. Component capabilities describe what the component can technically support. Whether the line achieves the intended protection also depends on configuration, use and operation. A testable acceptance question is therefore: Can the plant demonstrate that a maintenance authorisation has expired and that the approved project can be restored on the intended system? A modification, supplier change or newly reported vulnerability may change the assumptions. The operator records who will reassess them and approve the remaining risk. [6], [7]
Safety and security meet where digital events can influence a protective function. Machinery safety asks, for example, how hazardous movements are prevented while a person can access the work area. Industrial security also asks which changes to software, parameters or access might undermine that protection. Even when no person is endangered, the same plant may suffer serious losses in quality and delivery capability. Conversely, machine accidents can cause major operational consequences. The difference lies in the protection goals and the ways harm can occur.