Choosing an Industrial Switch: How It Differs from an Office Switch

The temperature, vibration, power supply and electromagnetic environment inside the cabinet must be compared with the technical limits of the switch to be used. Temperature, power, ring protocol, IGMP, management and lifecycle criteria for choosing an industrial switch.
Technical summary
Technology
Industrial switches, MRP, IGMP snooping, PoE
Topic
Network hardware selection
Level
Beginner–intermediate
Prerequisites
  • Information on the installation environment (cabinet, temperature, vibration)
  • Devices and protocols to be connected
In this article8 sections

In summer, the inside of a closed electrical cabinet is noticeably hotter than the surrounding environment. The temperature, vibration, power supply and electromagnetic environment inside the cabinet must be compared with the technical limits of the switch to be used.

In terms of its basic function, an industrial switch does the same job as an office switch: it forwards Ethernet frames to the right port. The difference is that it can do this in an electrical cabinet, next to a vibrating machine, at high temperatures and without interruption for years, and that it supports production protocols and redundancy needs. The criteria below explain, in order, where to look in the datasheet during selection.

1. Environmental resistance

  • Operating temperature: A wide temperature range for in-cabinet and field applications
  • Vibration and shock: Mounting on or near machines
  • Electromagnetic compatibility: Stable operation near drives, motors and welding machines
  • Dust and moisture: A suitable IP protection rating for use outside a cabinet
  • Fanless design: Cooling without moving parts

Pay attention to two things when reading the temperature value in the datasheet. First, the value is the temperature of the air around the device; inside a cabinet, next to drives and power supplies, this is higher than outside. Second, on some models the wide temperature range is a separate variant determined by the order code. If the standard variant is ordered and the wide range is assumed, the device may operate outside its technical limits. Environmental class, ring protocol support and a redundant power input also vary from model to model within the same product family; verify them in the datasheet by order code.

For vibration and shock resistance, manufacturers usually refer to the test methods in the IEC 60068-2 series. At special sites such as substations, railways or explosive atmospheres, the relevant industry standard (e.g. IEC 61850-3 for substations, ATEX/IECEx for explosive atmospheres) goes at the top of the selection criteria. At these sites, the “industrial” label alone is not enough. The certificate is checked.

2. Mounting and power

Industrial switches are usually DIN-rail mounted and powered with 24 V DC. In critical applications, a redundant power input and a relay contact that raises an alarm when power is lost are preferred. If cameras, handheld terminals or access points are to be powered, the PoE budget and DC supply capacity must be calculated together.

A redundant power input means the two inputs are fed from two separate sources. A common mistake is bridging both inputs from the same terminal of the same power supply. In that case, the power is not redundant. The fault relay contact also needs to be connected to the PLC or monitoring system; an unconnected alarm contact produces no alert.

For PoE, look not only at the power per port but at the total budget and input voltage. Under the standard, PoE output is in the range of roughly 44–57 V DC; models with a 24 V input generate this with an internal converter, and on some models the total PoE budget varies with the input voltage. The capacity of the 24 V power supply in the cabinet must also be calculated including the switch’s PoE load. If the cameras are heated (cold environment) models, the power drawn rises further.

3. Redundancy

For communication to continue over a backup path when a cable breaks in a ring topology, the switch must support the redundancy protocol in use. MRP is used in PROFINET environments, and RSTP or vendor-specific fast ring protocols in general networks. All devices in the ring must support the same protocol.

Mixing protocols leads to long faults in the field. When a switch from another manufacturer is added to a ring built with one manufacturer’s fast ring protocol, either the ring does not work or a loop forms. In a multi-vendor environment, preferring standard protocols (MRP, RSTP) makes future expansion easier. In applications with zero tolerance for interruption, PRP and HSR, defined in IEC 62439-3, send each frame over two paths at the same time, so there is no switchover time for a single failure covered by the design; however, these protocols require suitable hardware.

If RSTP is to be used, assign the root bridge priority manually. With the default setting, root selection is left to the MAC address, and the oldest, weakest switch on site may become the root. Recovery time also depends on the size of the topology; on long rings, the time must be compared with the limit the controller can tolerate.

4. Industrial protocol support

  • PROFINET compatibility and the switch’s ability to provide diagnostic information as a PROFINET device
  • IGMP snooping for multicast traffic in EtherNet/IP environments
  • QoS prioritisation for real-time traffic
  • Time synchronisation (PTP) support if needed

In EtherNet/IP, IGMP snooping alone is not enough; there must be an IGMP querier on the network. Without a querier, switches cannot keep the multicast membership table up to date, and the behaviour varies by model: traffic either floods all ports or is cut off unexpectedly. Defining one querier per VLAN is an easily overlooked setting. For design mistakes on the PROFINET side, see the article common mistakes in PROFINET network design.

5. Management and monitoring

A managed switch provides VLANs, port security, monitoring via SNMP, port error counters and log forwarding. These features are needed for both segmentation and fault finding. Being able to back up the configuration and load it quickly onto a replacement device (e.g. memory card or central management) shortens downtime.

Think about the moment of failure: night shift, a switch removed, a spare to be fitted in its place. If the configuration is on a memory card, the card is inserted into the new device and it starts up with the settings. If there is no card and the last backup is unknown, someone has to remember the VLANs, ring role and port settings by heart.

On the monitoring side, switches should send logs to a central syslog server and have their clocks synchronised with NTP. Logs from a switch with the wrong time do not line up with other systems’ logs during an incident investigation.

6. Security features

  • Changing default passwords and role-based permissions
  • Secure management protocols (HTTPS, SSH, SNMPv3); disabling insecure protocols
  • Disabling unused ports, MAC-based port security
  • The manufacturer’s policy for publishing security updates

IEC 62443-4-2 offers a reference for device-level security requirements; some manufacturers certify their products against this part. Even if certification is not required, whether the manufacturer publishes security advisories and how quickly it releases firmware when a vulnerability is found are selection criteria. For models that ship with Telnet and HTTP enabled from the factory, a “disable insecure services” item should be added to the commissioning checklist. For the concepts of zones and conduits, the article the zone and conduit approach in IEC 62443 frames these decisions. Address the switch and backup topology decision together with the distance, routing and testing requirements in the structured cabling and fiber infrastructure design.

7. Lifecycle

Production equipment stays in the field for many years. How long the chosen model will be manufactured, the availability of spare parts and the duration of firmware support should be part of the purchasing decision. Standardising on a small number of models across the plant is an advantage in terms of spare parts and know-how.

In practical terms, standardisation means the spare switch on the shelf can be fitted at any location in the plant. In a plant using five different brands and ten different models, the spare stock is either very expensive or incomplete. Follow the manufacturer’s end-of-life announcements and enter the dates into your own inventory; end-of-support dates vary from model to model, and current information should be verified on the manufacturer’s site.

Frequently asked questions

Why are industrial switches more expensive?

Requirements such as a wide temperature range, vibration and electromagnetic resistance, redundant power, DIN-rail mounting, long product life and industrial protocol support increase the cost. In return, the risk of downtime and the frequency of replacement decrease.

Is an unmanaged industrial switch enough?

It can be enough in small, isolated in-machine networks. At line and plant level, a managed switch should be preferred because segmentation, diagnostics and redundancy are needed.

Support

Contact us for support or a discovery session on this topic.

Contact form
[email protected]

Eretek Industrial

PLC, PROFINET/PROFIBUS and SCADA communication, IT/OT network segmentation and industrial field connectivity. We install and monitor without stopping the production line.

Knowledge Center

Other articles