The cybersecurity gap nobody's talking about in manufacturing
For decades a factory network was simple. Machines sat on their own, wired to their own controllers, and few outsiders could reach them. That isolation was a kind of security. Then plants began connecting machines to dashboards, planners and vendors, and the isolation went away. Production improved. The attack surface widened with it.
For decades a factory network was simple. Machines sat on their own, wired to their own controllers, and few outsiders could reach them. That isolation was a kind of security. Then plants began connecting machines to dashboards, planners and vendors, and the isolation went away. Production improved. The attack surface widened with it.
Connecting machines helps production and widens the attack surface. Most manufacturers have not yet caught up with the second half of that sentence. This article sets out four common gaps and the first steps to close them.
01.Why factories are different from offices
Office IT protects data. Factory systems, usually called OT, protect physical processes. If an office server goes down for an hour, people wait. If a controller misbehaves, a line stops, product is scrapped or someone is put at risk.
That changes the rules. You cannot patch a PLC on a Tuesday afternoon without thinking about the line. Many devices run for twenty years. Some cannot take a security update at all. Security measures that suit a laptop can break a production process. So OT security has to be designed with the plant team, not handed down by IT alone.
02.Gap 1: Old equipment that cannot be patched
Plants often run machines and controllers made long before security was a design concern. Their software may be unsupported and have known weaknesses. Replacing them is expensive and sometimes impossible within a budget cycle.
The practical answer is to stop them being reachable. Put them in their own network zone, allow only the traffic they need, and monitor what they do. You cannot fix the device, but you can limit who talks to it.
03.Gap 2: Flat networks linking office and factory
In many plants the office network and the production network are one and the same. A phishing email that lands on an office laptop then has a path to the machines.
Segmentation addresses this. Split the network into zones, such as office, plant servers and individual cells, with a controlled point between each. Allow only the connections that are needed, and block the rest by default. If an attacker gets into one zone, they should not walk straight into the next.
04.Gap 3: Remote vendor access without controls
Machine suppliers often need remote access to diagnose faults. The common setup is a shared account, a permanent remote tool or a modem left switched on. These are convenient and dangerous.
Better practice is to give each vendor person a named account, require multi-factor authentication, and open access only for the duration of a job. Route sessions through a single gateway that records what was done. Review the list of who has access every quarter, and remove accounts that are no longer needed.
05.Gap 4: No monitoring of OT traffic
Many plants watch their office network and have no view into the factory network. If something odd is happening, such as a new device, an unexpected command or a scan across the network, nobody knows.
OT monitoring is passive. It listens to traffic and builds a picture of normal behaviour: which devices exist and how they talk. It then alerts when something changes. It does not need to touch the controllers, which makes it safe to deploy.
06.First steps
The first steps are clear and do not need a large budget.
- Inventory assets. List every machine, controller, switch and remote connection. You cannot protect what you cannot see.
- Segment networks. Separate office from plant, and the plant into zones. Start with the most critical lines.
- Control remote access. Named accounts, multi-factor authentication, time-limited sessions and logging.
- Monitor. Passive monitoring of OT traffic with alerts that the plant and IT teams both understand.
Do them in that order. Each step makes the next easier and cheaper.
07.What a bad day looks like
It helps to make the risk concrete. Imagine a plant where the office and production networks are joined. A staff member opens an email attachment on an office laptop. The malware spreads through the shared network and reaches a server that talks to the production machines. Lines stop while engineers work out what is infected. Controller programs are not backed up, so restoring them takes days. Orders slip, and customers ask questions.
None of this needs a sophisticated attacker. It needs a flat network, an unpatched system and no plan. Every one of the four gaps above makes this story more likely, and every first step makes it less so.
08.Working with plant teams
Security measures succeed or fail on the shop floor. Engineers value uptime above all, and rightly. A new rule that interrupts a line will be bypassed. So involve the plant team early. Explain what each control does in terms they care about: fewer surprises, faster diagnosis, clear records of who touched what.
Agree change windows in advance. Test new firewall rules in monitor mode first. Keep a documented way to switch a control off in an emergency, with a record of who did it. A control the plant trusts is a control that stays in place.
09.Vendors and contracts
Remote access is partly a contract matter. Ask equipment suppliers how they connect, what accounts they use and whether they can work through your gateway. Put expectations in the support agreement: named users, multi-factor authentication, notice before sessions and a log you can read. Suppliers who cannot meet reasonable terms are a risk you should know about before an incident, not after.
10.Starting small in a real plant
A realistic first quarter looks like this. In month one, walk the plant and build the asset list. Expect to find devices nobody remembers installing. In month two, separate the office network from the plant network and close any direct paths. In month three, move vendor access onto a single controlled gateway and switch on passive monitoring for the most critical line.
At the end of three months you will not be finished, but you will know what you have, who can reach it and when something changes. That is a very different place from where most plants begin, and it was reached without replacing a single machine.
11.Metrics that help the board
Security spending is easier to defend with plain measures. Count the assets you have identified against those you estimate exist. Count the zones in the plant network and the rules between them. Count vendor accounts, and how many are named and time-limited. Track the time between an unusual event on the plant network and someone looking at it. These figures are modest, and they show direction without claiming certainty.
12.Common mistakes
- Treating OT like office IT. Scanning tools and patch routines can interrupt a line.
- Leaving it to IT alone. Plant staff know which connections matter and what a safe change looks like.
- Buying tools before knowing the assets. A monitoring product without an inventory produces noise.
- Forgetting backups. Keep tested copies of controller programs and configurations so you can recover after an incident.
- No plan for an incident. Decide in advance who stops what, and who calls whom.
13.Closing
The gap in manufacturing security is not mysterious. It comes from old equipment, flat networks, loose vendor access and silence on the factory network. Each can be reduced with practical steps. If you are connecting your plant and want a sound design for it, Brandsmashers works on OT/IT connectivity and security, and can review your current set-up with you.
NEXT STEP
Need help with this?
KEEP READING
