Industrial Ethernet: managed or unmanaged switches?
|PTO offers specifications, profiles, guidelines, installation information, brochures, technical write-ups, white papers, and more.|
Just because you can use an unmanaged switch does not mean you should use an unmanaged switch. It has been said that you’ll either use a managed switch or wish you had, explains Carl Henning, deputy director, PTO, Profibus and Profinet North America .
Here are additional Henning Ethernet observations:
The logic is that an unmanaged switch has diagnostic information consisting of blinking LEDs. If your line goes down due to a switch failure and you can tolerate the downtime to go find blinking lights, an unmanaged switch might be the right choice for you.
If you can’t stand the downtime, then use a managed switch. Its onboard diagnostics can be read by many kinds of software from the IT world.
Use an SNMP (Simple Network Management Protocol) OPC Server to move diagnostic data from the switch to an HMI. Managed switches maintain a database of important switch activity.
The database is called an MIB (Management Information Base). For example, numerous retries on a particular switch port might indicate a network problem. Alerting maintenance through the HMI could prevent an unplanned shut down.
Of the three most popular Industrial Ethernet protocols, Profinet and Modbus/TCP can use either managed or unmanaged Ethernet switches; Ethernet/IP requires the use of a managed switch. Ethernet/IP needs a managed switch because it requires the IGMP (Internet Group Management Protocol) Snooping function which is not available in unmanaged switches. Ethernet/IP uses broadcasts to transmit IO data. To prevent this from overloading the network, IGMP Snooping learns where broadcasts messages are going and restricts them to only the interested nodes. This prevents the network from being overloaded with traffic.
Some Ethernet switches can also act as IO devices. For example, a switch that uses Profinet to access the MIB can move diagnostic data to a controller in addition to an HMI. The controller could then react and respond to the condition rather than just creating a notification on an HMI screen.
This ability to use information that is already there overcomes one of the major concerns of control engineers and maintenance staff: “How can I diagnose and fix network problems? Do I need special equipment? My maintenance staff doesn’t carry laptops.” Downtime can be minimized by bringing this information into systems that automation folks are already familiar with (HMIs and PLCs) and are already installed.
– A related white paper that helps with some of those decisions while pointing out differences between Ethernet for IT and automation domains (PDF) ; and
: Industrial Ethernet protocols can be standard and unmodified, or modified to be industrially hardened. Perspectives from some protocol proponents clarify needs for various implementations.
– Edited by Mark T. Hoske , editor in chief
Control Engineering News Desk
Register here to select your choice of free eNewsletters .