How to scale factory networking for UK digital transformation

A scalable factory network is one that lets production, engineering and business systems grow without making uptime, security or operational change harder to manage. For a UK manufacturer, that means designing the wired foundation, OT and IT boundaries, mobile connectivity, identity controls and monitoring as one connected system — then introducing change in phases that can be tested around production.

This guide explains how to approach that design without treating every factory as a generic office or data centre.

Why is factory networking harder to scale than an office network?

A factory network carries more than user traffic. It may connect PLCs and PACs, robotics, automated guided vehicles, machine-vision systems, HMIs, engineering workstations, sensors, maintenance equipment and business applications. These systems have different availability, latency, security and lifecycle requirements.

The physical environment adds another layer of complexity. Metal racking, machinery, moving vehicles, interference and extended temperature ranges can affect wireless performance and equipment selection. A design that works in an empty building or office corridor may fail when production is running, routes are occupied and mobile equipment is moving.

The business consequence is more important than the technical detail. If a network change interrupts a production line, delays quality data or prevents engineers from reaching the systems they need, the cost is measured in operational disruption rather than merely in lost connectivity. Scalability therefore means predictable change: adding a line, site, device type or digital service without creating an untested dependency.

What should a scalable factory network help the business achieve?

Start with production flows rather than a preferred product or topology. Identify which processes cannot stop, which systems need real-time or near-real-time access, and which data must move between the plant and enterprise platforms. This creates an acceptance test for the network design.

The goal is not to make every connection equally fast or every component industrial-grade. It is to provide the right level of resilience, isolation and performance for each zone. A critical control path may need a wired connection and carefully tested failover. A visitor may need simple, isolated wireless access. A sensor fleet may need secure, scalable onboarding without being able to reach production systems.

This outcome-led approach also prevents over-investment. Industrial hardware, specialist wireless and additional redundancy have a place, but only where the environment or business consequence justifies them. Spending more on equipment everywhere is not the same as designing a more resilient network.

How should you separate factory OT, enterprise IT and connected devices?

A scalable design should not rely on one flat network with trust implied by physical location. Production, engineering, corporate, supplier, guest and IIoT traffic should have deliberately defined boundaries, permissions and routes.

Macro-segmentation can separate major environments, while more granular policy can control communication between devices, users and services within them. In a Cisco-oriented architecture, that can involve VRFs, overlays, security groups and identity-led policy, with an Industrial DMZ providing a controlled boundary between plant OT and enterprise IT. The specific implementation should follow the site’s systems, protocols and operational constraints.

Identity is part of the architecture, not a later security add-on. 802.1X, certificate-based authentication, network access control and posture checking can help distinguish employees, contractors, suppliers, managed devices and unknown equipment. Wireless access should apply the same principles, using appropriate enterprise authentication and isolating guests and BYOD from operational systems.

The practical test is simple: can the organisation explain who or what is allowed to communicate, why that access is needed, and how it would be removed or reviewed? If not, the network may be connected, but it is not yet governed at scale.

Which factory wireless technology should you use?

Wired connectivity should remain the starting point for safety-critical or fixed systems wherever cabling is feasible. Wireless is valuable when mobility, difficult cabling routes or operational flexibility create a genuine business benefit.

Wi-Fi 6 or 6E can suit workforce devices, HMIs and many general-purpose workloads when coverage, capacity and roaming are designed around the real environment. Mission-critical mobile links may need a different approach. Ultra Reliable Wireless Backhaul or private cellular can be considered for equipment such as AGVs and AMRs where bounded latency, resilient paths or indoor-to-outdoor continuity matter.

That does not mean every moving asset needs specialist wireless. The right decision depends on the application’s tolerance for delay and interruption, the route it follows, the available wired alternatives and the consequence of failure. A reliable design validates coverage at client height, tests hand-offs along actual routes and accounts for racking, machinery and seasonal changes in the environment.

A survey carried out before the factory is operational can be useful, but it is not acceptance evidence on its own. Route testing, failover testing and application-level validation are what turn a coverage plan into a production-ready design.

How do you build visibility without creating another dashboard problem?

Digital factory initiatives often promise better visibility, but the network underneath can become fragmented as each platform adds its own connectivity, monitoring or security assumptions. The result is a dashboard that looks useful until a link fails, a policy blocks traffic or a new device behaves differently from the original design.

Monitoring should join infrastructure health with the production context that matters to the business. At a minimum, teams need a dependable view of device health, link utilisation, wireless performance, authentication events, configuration changes and critical paths. They also need clear ownership when an alert crosses the boundary between IT, OT, a supplier and a managed service provider.

Standardised configurations, version control, automated backups and tested rollback can reduce avoidable variation across sites. Templates and infrastructure-as-code may help where the environment is sufficiently consistent, but automation should be introduced with change controls and validation. Automating an unclear or unsafe design only makes mistakes repeatable.

The measure of good observability is not the number of metrics collected. It is whether an engineer can identify what changed, what is affected and what safe action is available before a local issue becomes a production problem.

How can a factory migrate without disrupting production?

A phased migration is usually more credible than an overnight replacement. Begin with an inventory of devices, applications, protocols and dependencies. Include Modbus, PROFINET, EtherNet/IP and OPC UA where relevant, but do not assume that a protocol inventory alone explains the operational risk. Map the flows and the people or suppliers who support them.

Next, define the target boundaries and the evidence needed to accept them. That might include redundancy and failover tests, route testing for mobile equipment, authentication checks, controlled supplier access and confirmation that production data reaches the required enterprise or analytics service.

A pilot should represent the difficult parts of the environment, not only the easiest cabinet or cleanest wireless zone. Keep a tested rollback route and schedule changes around production constraints. Once the pilot is accepted, repeat a standard site or zone pattern while allowing for genuine local differences.

This approach protects existing investment. Legacy equipment can remain in service where it is safe and supportable, while its access and communication paths are made more controlled. Replacement should be driven by operational risk, lifecycle position and the value of the change — not by the desire to make every component new at once.

How we think about scalable factory networking

We begin with the site and the production outcome. What is being made? Which processes must remain available? Where is IP-rated or extended-temperature equipment actually required? Which protocols cross the OT and IT boundary? Which mobile assets need deterministic connectivity, and which can use standard wireless?

From there, the design should join five disciplines: a resilient wired foundation, deliberate OT and IT segmentation, identity-led access, validated mobility and actionable monitoring. The architecture may use different technologies across different sites, but the policy, documentation and acceptance criteria should be repeatable.

We also believe the customer should own the underlying infrastructure rather than allow each MES or IIoT platform to bolt on its own network. That makes future digital services easier to govern and helps avoid dashboards that look impressive until connectivity, segmentation or monitoring breaks under real production conditions.

Frequently asked questions

Is SD-WAN enough to make a factory network scalable?

No. SD-WAN can help with multi-site connectivity, traffic policy and cloud access, but it does not solve plant-floor wireless, OT and IT boundaries, industrial hardware requirements or application-level failover. It should be considered as one part of a wider architecture.

Should every device in a factory use an industrial network switch?

Not necessarily. Industrial equipment is justified where environmental conditions, installation location, lifecycle or operational consequence require it. Specifying it everywhere can add cost and complexity without improving the outcome.

Is wireless suitable for production equipment?

It can be, depending on the application and its tolerance for delay or interruption. Fixed and safety-critical systems should generally use wired connectivity where feasible. Mobile equipment needs route-based testing and a technology choice based on the workload, not simply on coverage.

How do you secure supplier and contractor access?

Use identity-based access, strong authentication, device and posture controls where practical, least-privilege permissions, time-bounded access and a controlled path into the relevant environment. Access should be logged, reviewed and removed when the work ends.

What should be measured after a network upgrade?

Measure the outcomes agreed at the start: availability of critical paths, failover behaviour, application performance, roaming along real routes, authentication success, incident diagnosis time and the ability to add approved devices or services safely.

In short

Factory networking scales when it is designed around production flows rather than device counts. Connect the wired plant, enterprise IT, mobility, identity and monitoring layers; segment deliberately; test in the live environment; and migrate in controlled stages. The right design may use SD-WAN, Wi-Fi 6, specialist wireless or private cellular, but only where each supports a clear operational outcome.

If you are considering a factory network change, a practical conversation about production constraints and acceptance criteria is often the best place to start.