An industrial Ethernet network is vastly different from an office Ethernet network in several key ways, and the key to optimizing your industrial network in light of these differences, is hands-on training.
First of all, the environment in industrial applications can degrade the actual cable itself. Some cable manufacturers actually rate their cables’ ability to withstand these environmental factors. They use the acronym MICE, and rate the cable as appropriate for one of three environments: M1I1C1E1 for office environments, M2I2C2E2 for light industrial environments, and M3I3C3E3 for industrial environments. The letters actually stand for: Mechanical factors such as shock and vibration, Ingress from moisture, Climatic factors such as temperature and sunlight, and Electromagnetic interference such as noise caused by inductive loads, welders, variable frequency drives, etc. Other cable vendors observe the recommendations of ODVA and offer cables that are ODVA compliant.
Secondly, industrial Ethernet networks can have a high amount of multicast traffic. In the early years of Ethernet hubs were used to link devices. The problem is that information coming into one port of a hub was redirected to all of the other ports on the hub. With the advent of switches, unicast traffic was now directed to only the port for the intended recipient device. This is true for both managed and unmanaged switches: they both handle unicast traffic well. The problem for the unmanaged switch comes when you encounter multicast traffic. Since an unmanaged switch does not employ IGMP Snooping (Internet Group Management Protocol), the switch does not know what to do with multicast traffic. It starts acting like the old hubs: it directs all multicast traffic to all ports. With a managed switch and with IGMP Snooping turned on, the switch knows exactly where to send this multicast traffic and directs it only to the intended recipients. Multicast traffic can be anything from produced tags to input modules configured for multicast. These can be very common in industrial applications using PLCs.
Thirdly, we now have tools available in many switches and routers to prioritize the traffic on an Ethernet network. This becomes especially important when you have high-speed applications, motion applications, or time synchronization applications. In the past all Ethernet data was equal. The feedback coming from a servo drive had to wait just as long as a person trying to get online with a PLC. Now many automation vendors are marking their data with priority codes. Allen-Bradley marks their data in layer three with DSCP markings, and Siemens uses layer two markings with PCP marks, for instance (a VLAN tagging mechanism). In either case, if your switch or your routers are not configured properly to recognize and use these priority codes, you are not taking advantage of the QoS feature that could help get your important data through first (Quality of Service).
Only through proper training can you learn not only what the key issues are but also how to properly deploy your hardware to fully optimize your network. Balluff offers hands-on training with actual automation equipment, switches, and routers to help you do just that. You can learn more about the courses Balluff has to offer at www.balluff.us/training.