Objectives
This Technical Note provides some inside information on how device addressing is organized within PROFINET.
Introduction
When comparing an office Ethernet environment with what is common on the factory floor there are huge differences. Not only on the physical side, but also in how the addressing and the traffic work. The content below examines how this is handled by PROFINET.
Content
When a computer is first booted up (usually first thing in the morning), it gets itself an IP-address from the networks DHCP-server. The IP-address may look something like 192.168.250.10, for as long as the computer is connected. On the next bootup, it may well be a different address.
Getting a different address each time is not a problem when the computer is transferring network traffic to a known source, like the fileserver or the web. But with a machine it is essential that the device it is trying to communicate with is the correct one. It, therefore needs some form of consistency. For example, it would not be acceptable in practice for a command to one day open a valve whilst on the next day the same command is directed at starting a motor. There is also the issue of device replacement. If a device breaks down it must be easy to replace and reconfiguration must be limited to a minimum.
Background as to how PROFINET communication functions.
In all industrial communication protocols there are two types of communication:
1. Configuration and diagnostics.
2. I/O data exchange.
PROFINET uses normal TCP/IP communication for the configuration and diagnostics tasks. Therefore each device needs to have a TCP/IP address.
The I/O data exchange does not use TCP/IP communication. It is very low level Ethernet communication. Whole TCP/IP is bypassed and PROFINET works directly on the Ethernet controller. This is done so that the PROFINET-IO frame has less overhead in the protocol handling and becomes smaller. This increases the throughput because less computing power is needed.
In order to make a device unique on the network and to be able to use TCP/IP communication with it, the following is recommended. Normal industrial practice is to use a couple of rotary switches on the device to set the address. But in PROFINET the number of devices is unlimited. This means it would require an unlimited number of rotary switches.
Therefore PROFINET implemented the Name mechanism. To make a device unique it is given a name, a string that is loaded in the device. A PC-software tool is normally used, incorporated in the Network configuration tools. It sends out a query message on the network to which all PROFINET devices react. In the tool a list appears of all discovered devices with their MAC-ID, their default name and a default IP-address. The MAC-ID is the low-level Ethernet-address of a device.
There may well be several devices, e.g. SmartSlice I/O stations, which have the same default name out-of-the-box.
To find which device is where, there is a “blink” option. Select one of the devices found and activate the “blink” function. Then the device will start to blink its LEDs. At this point it is necessary to confirm the position of the blinking device. With the position is a name or code associated with it and that is what you program into the device. It is now uniquely named.
Now enter the names into the configuration of the PROFINET IO-controller. Plus an IP-address that is needed to do configuration or any other TCP/IP communication (FTP, WEB, etc.).
The PROFINET IO-controller knows the names of the devices it needs to communicate with. When a PROFINET IO-controller starts, it looks out on the network for a device with a particular name. The device with this name replies and the PROFINET IO-controller then issues the device its IP-address. Now the device is ready to receive its parameters and configuration and the IO-control cycle starts.
What happens when a machine is faulty and a part of it needs to be replaced? This would be ok, if the replacement device wasn't the part that held the stored name.
Getting a different address each time is not a problem when the computer is transferring network traffic to a known source, like the fileserver or the web. But with a machine it is essential that the device it is trying to communicate with is the correct one. It, therefore needs some form of consistency. For example, it would not be acceptable in practice for a command to one day open a valve whilst on the next day the same command is directed at starting a motor. There is also the issue of device replacement. If a device breaks down it must be easy to replace and reconfiguration must be limited to a minimum.
Background as to how PROFINET communication functions.
In all industrial communication protocols there are two types of communication:
1. Configuration and diagnostics.
2. I/O data exchange.
PROFINET uses normal TCP/IP communication for the configuration and diagnostics tasks. Therefore each device needs to have a TCP/IP address.
The I/O data exchange does not use TCP/IP communication. It is very low level Ethernet communication. Whole TCP/IP is bypassed and PROFINET works directly on the Ethernet controller. This is done so that the PROFINET-IO frame has less overhead in the protocol handling and becomes smaller. This increases the throughput because less computing power is needed.
In order to make a device unique on the network and to be able to use TCP/IP communication with it, the following is recommended. Normal industrial practice is to use a couple of rotary switches on the device to set the address. But in PROFINET the number of devices is unlimited. This means it would require an unlimited number of rotary switches.
Therefore PROFINET implemented the Name mechanism. To make a device unique it is given a name, a string that is loaded in the device. A PC-software tool is normally used, incorporated in the Network configuration tools. It sends out a query message on the network to which all PROFINET devices react. In the tool a list appears of all discovered devices with their MAC-ID, their default name and a default IP-address. The MAC-ID is the low-level Ethernet-address of a device.
There may well be several devices, e.g. SmartSlice I/O stations, which have the same default name out-of-the-box.
To find which device is where, there is a “blink” option. Select one of the devices found and activate the “blink” function. Then the device will start to blink its LEDs. At this point it is necessary to confirm the position of the blinking device. With the position is a name or code associated with it and that is what you program into the device. It is now uniquely named.
Now enter the names into the configuration of the PROFINET IO-controller. Plus an IP-address that is needed to do configuration or any other TCP/IP communication (FTP, WEB, etc.).
The PROFINET IO-controller knows the names of the devices it needs to communicate with. When a PROFINET IO-controller starts, it looks out on the network for a device with a particular name. The device with this name replies and the PROFINET IO-controller then issues the device its IP-address. Now the device is ready to receive its parameters and configuration and the IO-control cycle starts.
What happens when a machine is faulty and a part of it needs to be replaced? This would be ok, if the replacement device wasn't the part that held the stored name.
Summary
Our opinion is that the name storage should be part of the functioning device and be recognizable as such. For our SmartSlice system the name storage function is put in a memory end-plate, called GRT1-END-M. When the name is first assigned to a SmartSlice station, the name is stored in the bus coupler, and automatically copied into the memory end-plate.
The faulty buscoupler is replaced with a spare one. The buscoupler will notice there is a GRT1-END-M with a stored name and will start using that name. The PROFINET IO-controller will find the buscoupler with the correct name and starts communication with it. Problem solved.
SmartSlice’s automatic name backup and restore make device replacement quite easy in PROFINET.
