Avoid network failures utilising Profinet analysis
Advantages of using a passive diagnostic tool for Profinet analysis
What is Profinet?
Profinet (Process Field Network) is the open Industrial Ethernet standard for data communication in the automation industry. It is designed for cyclic data exchange between the controller (= I/O controller) and the field devices (= I/O devices). For this purpose, Profinet uses I/O Ethernet-based protocols.
Profinet has three major architectural components :
- I/O controller: The controller that controls the automation task.
- I/O devices: Field devices that are controlled and managed by the controller. Field devices can consist of several modules and submodules. The submodules contain input and output signals to the process.
- I/O supervisor: The engineering tool to parameterize and diagnose individual I/O devices.
The process flow
The controller sends data to the I/O devices. These signals are forwarded by switches. The process can be compared to a priority buffer: first come, first served (or what is processed first). If the priority buffer is full, then the data is discarded. The data rate of the devices is 100 Mbit/s.
Switches operate in "store and forward" mode, in which the data is first checked for CRC errors before they are forwarded. If the data is compromised, it is deleted. Therefore, time delays (jitter) can occur with each switch. Traffic is most congested between the controller and the first switch. This can put additional strain on the network.
In addition, the data exchange is hampered by larger packets from TCP or IP traffic. This can cause queues on switches when there is a high flow of traffic. If the buffer memory of the switches is full, excess data in the queues are deleted. This can lead to packet losses and breakdowns. To prevent possible errors at early stages, the network should be permanently controlled or monitored. Network diagnostic tools are suitable for this purpose. However, consideration should be given to the additional load that an active diagnostic tool can place on the network, therefore is recommend that passive network tools are used and why passive tooling is recommended.
The Allegro Network Multimeter provides high granularity for Profinet troubleshooting and control. Packet-based analysis and statistics are available for communication relationships, top talkers, bandwidth consumption, jitter, frames, Profinet errors, alarms, and more.
Profinet traffic is recorded and can be viewed in real-time from both current and past traffic. In doing so, anomalies such as jitter or errors can be isolated by selecting a peak in the graph using 'click and drag'. This helps narrow down the time period for troubleshooting and get pre-filtered pcap files with Profinet traffic for MAC or IP when needed.
The incident function can be used to set up automatic messages when specified thresholds for Profinet are exceeded. Detailed instructions on how to find errors in Profinet in a time-saving way are available here.
Why is the Profinet analysis not done directly with Wireshark?
For a global overview of sporadic errors (what is happening in the system?), the Allegro Network Multimeter is more suitable. It is possible to examine the live traffic and the retrospective results. Wireshark is a useful tool when the display of each packet and the included logs are required and when the data of an installation is to be examined over a short period of time. However, if the disruption occurred five hours ago (rather than minutes) and it is not clear where it occurred, a large pcap would first need to be recorded and filtered by hand. This process is much faster with the Allegro Network Multimeter. It offers a time-saving analysis with many more features.
As always, more information about the features of the Allegro Network Multimeter can be found in the Product Wiki.
Use Case from Video
Product Manager and Managing Director of Allegro Packets, Klaus Degner, gives you an insight into Profinet analysis with the Allegro Network Multimeter in this video (in German).
The video uses a custom trace to show how you can use pcaps to analyze Profinet. In the Allegro Network Multimeter at 'Generic' you can open the pcap analysis, put the pcap in and analyze it. So simple.
For the Allegro Network Multimeter in the video, the process is started offline. The link is not broken during the parallel offline analysis. As soon as the analysis is started, it is visible that there was traffic in the 2.9 megabytes. All statistics can be displayed at 'Applications' → 'Profinet statistics'.
Profinet is an industrial protocol with three services. There are:
- synchronous messages, in which timed data are retrieved
- asynchronous messages, in which there is measurement data on request and
- alarms
In the dashboard statistics, you can quickly look into the individual services: Which device is there, what are the global statistics, what is happening in Profinet, what is the jitter right now for these cyclic messages, etc.?
Installations have often found that Profinet itself is not the problem, but that other network services are causing interference. There is the function 'Graphs' to see the Profinet traffic compared to the total traffic. In the global statistics the different errors can be viewed. How many errors were there? How many alarms were sent with high priority? What was the maximum jitter? You can go into more detail on each question and look at all the devices that were in the Profinet and which one had how much traffic. This works both for the current traffic and retrospectively. Each time interval can be selected, the required time period zoomed in and extracted as a pcap.
Conclusion
Most errors as well as an overload in Profinet rarely happen due to Profinet itself, but more often due to other interfering services. Regular control of Profinet with a passive network analysis tool is essential to ensure production processes without delays.
