CourseMonster

Real-time and Predictable Latency Capabilities in Software-defined PLC - Course Monster Blog

Written by Marbenz Antonio | 17/08/2022 4:30:30 AM

The software-defined industrial control system represents a big opportunity for open source technology. A trend that cuts across industries is moving away from custom-made specialized hardware (like PLC devices) and toward standard hardware with software alone. This movement, emphasizing real-time and predictable latency capabilities, has now entered the industrial and manufacturing industry. Like telcos’ use of network function virtualization (NFV) and open radio access networks, industrial enterprises are making a similar transition (ORAN).

However, Real-time and Predictable Latency Capabilities are necessary for this trend to be implemented successfully. When an external event occurs, such as a robot about to hit a wall, the controller must act quickly to stop the robot before it makes contact. Please be aware that we do not use the phrase “real-time,” as this concept calls for specialist Real-Time Operating Systems (RTOS). Predictable latency is the better alternative.

Red Hat Enterprise Linux (RHEL) for Real-Time, which provides the foundation for the Red Hat OpenShift Performance Addon Operator, is one of Red Hat’s responses to this expanding industrial trend (PAO). Although it was created for the telco industry, it may also be highly effectively used in operational technologies and industrial control systems.

The outcome of joint testing and assessment efforts between our two businesses, we presented Digital Transformation at the Edge with Red Hat and Intel Edge Controls for Industrial with Intel.

Let’s take a look at the data:

Red Hat OpenShift on bare metal with the Performance Add-On Operator and the RHEL Real-Time kernel is used for this test (RT-Kernel). Maximum latency (the time between an event and a response) is between 47 and 33 seconds. Except for turning on Intel Cache Allocation Technology, this is very much out of the box on ordinary hardware. With PLC cycle rates in the 10 msec range, this is already sufficient for many industrial control system applications, leaving enough headroom for the actual program execution.

Want to know more about Red Hat OpenShift? Visit our course here.

But it gets even better:

This graph illustrates the benefits of the RHEL RT-Kernel that is deployed by the OpenShift Performance Add-On Operator. You’ll see that we changed the measurement from latency (how quickly the system responds) to jitter (how consistently the system responds, which is more important). It is around 57 msec on a default system (standard kernel), which is too long for the majority of control applications.

However, this decreases to 55 µsec simply by activating the OpenShift Performance Add-On Operator, with no further configuration. That represents a three-order-of-magnitude boost.

Let’s take a look at what’s possible when we add Intel capabilities to the equation:

The wide deviations on the left-hand side are caused by CPU cache misses. It takes a while to fetch the code or data from memory if it is not already in the cache. First and second-level caches may be affected by concurrent processes or other workloads on the same core. By allocating parts of the CPU cache to a process, Intel Cache Allocation Technology prevents this.

With a 76% decrease in maximum jitter, the previous minimum becomes the new maximum when compared to the new statistics.

Real-time and Predictable Latency Capabilities: Conclusion

The statistics demonstrate that in containerized settings, Real-time and Predictable Latency Capabilities workloads like industrial control systems may be operated on contemporary standard hardware. No need for long-term heavy fine tuning or specialized equipment. For the majority of use situations, the performance characteristics are sufficient out of the box.

Unlock new information about Real-time and Predictable Latency Capabilities. Visit us now.

Here at CourseMonster, we know how hard it may be to find the right time and funds for training. We provide effective training programs that enable you to select the training option that best meets the demands of your company.

For more information, please get in touch with one of our course advisers today or contact us at training@coursemonster.com