PREDICTIVE MAINTENANCE

Kenton Williston: Welcome to the IoT Chat, where we explore the trends that matter for consultants, systems integrators, and enterprises.  I’m Kenton Williston, the Editor-in-Chief of insight.tech. Every episode, we talk to a leading expert about the latest developments in the Internet of Things. Today, I’m discussing the new 12th Gen Intel^® Core^™ processors with Kenny Chang, Vice President of the System Product Business Unit at ASRock Industrial.

The new CPUs, which were formally named Alder Lake, were just announced at CES 2022. They pack a ton of cool features, like an all-new hybrid architecture, massively upgraded GPUs, and real-time capabilities. As one of the first companies to come to market with the new chips, ASRock has unique insights on these processors. So I’m really looking forward to hearing Kenny’s thoughts.

So with that, Kenny, I would like to welcome you to the podcast.

Kenny Chang: Hi, thank you for having me and it’s my honor to join the podcast.

Kenton Williston: Yeah, absolutely. I’m curious about your career. What have you done before your previous role at ASRock?

Kenny Chang: I was in charge of the product-development division as the vice president, and our product is major in medical devices for what we call IoM. That means Internet of Medical. And the major equipment we are developing is the front panel dictator. It is used for the X-ray system. That’s what I did before I joined ASRock Industrial.

Kenton Williston: So as I was just saying, there are so very many new features in the latest Intel core processors. So there are a lot of things we could spend our time talking about, but the first thing I would like to know is what features you are most excited about in the latest Intel core processors and why?

Kenny Chang: I think the most amazing feature is about the hybrid architecture, combining the performance core as well as the efficient core. I think that gives us the very flexible way to manage, especially in software-defined everything, we can adjust which core is doing what kind of jobs accordingly. And I think this is the major benefit when we adopt the Alder Lake-S processor into our products.

Kenton Williston: Yeah, that totally makes sense. And I should mention too that this podcast as well as the overall insight.tech program are produced by Intel, so of course, we have very specific reasons to want to talk about the latest Intel technologies. But having said that I agree the hybrid architecture is very interesting. This is something that’s become more popular in a variety of CPU designs and, just like you said, having both high performance and good efficiency is a really amazing combination.

You can combine these two things together and not have to give up low power to achieve high performance or vice versa. That’s very, very helpful. One of the things that Intel has been saying about these new processors is that they are revolutionary, which is, of course, a very strong claim and I’m wondering how you think they might revolutionize IoT applications.

Kenny Chang: The IoT application is very diversified and across various vertical applications such as the automation, automobile, smart city, energy, or even smart retail. We have so many software applications onto the end products. We are going to the microservice enabled by the containers. That means there’s so many containers running simultaneously on the one edge platform. Especially the edge is the most important. I think it is a major migration from the data center to decentralized. That means we have edge computing running on the local sites. They can reduce then enforce all the latency between the cloud and the devices. So edge is the best solution for this kind of situation.

Back to the 12th Generation Alder Lake processor here, we can embed the powerful processor into the  edge devices. So, with that, we need not only powerful, but also traceable architecture to deal with all the microservices or tasks. That’s the first one I would like to highlight here. And the second one is, we do lots of critical missions on the edge. The real time is really a method for every operation, and our 12th generation possessor Alder Lake-S featuring about the real-time controlling by the TSN and the TCC.

So that’s a good feature to make us to have more confidence: to make all tasks be occurred, synchronization in the real-time measure. This is the two major key benefits for the IoT application.

Kenton Williston: Yeah, I think those are all really good points. So, let’s see if I can summarize those and add hopefully something useful on top of it. The point you made about microservices and containers, I think is a very good one. It’s a very reflective of a big change in how edge computing is done. I hate to date myself like this, but I’ve been working in this space now for let’s see, I guess this was going to be 22 years this year. When I began my career, the things that you would find in what we’d call now, edge computing, which back then would be called embedded computing, were very specialized code. You had to have very specific knowledge to write for these devices. And now people are looking more and more to use the same kind of coding practices that you would find people using in the cloud.

And I think this is a good change because one thing that’s very good about it is it gives you so much more flexibility. Because it is important, I think, to move some things out of the cloud to the edge, but of course there’s always this back and forth. Sometimes things need to be decentralized, sometimes they need to be centralized. And I really like the way that with modern application development processes, you gain a lot of flexibility. The microservice can just run wherever it makes the most sense. But of course you need, in order for that to happen, you need to have a platform that will support running these microservices. And I think the older lake platform is a very good one for that.

Kenny Chang: Yeah, exactly.

Kenton Williston: And then you mentioned the importance of the hybrid architecture, and one of the things that’s important here is you can configure your system in such a way so that the less performance hungry microservices, and just in general the less performance hungry tasks, can run on the efficient cores, so that if you don’t actually need the high performance cores at the moment, you’re running very efficiently, very low power, which is important for all kinds of reasons. Obviously, if you have something that is battery powered, it’s very important not to draw too much power, but even if you have something that is plugged into the wall, in many situations, it’s very important not to have too high of thermals because then you start having much more complicated systems that are less reliable and have more moving parts and all these sort of things.

And of course, in addition to this overall trend of edge equipment looking more and more, at least from a software perspective, the same as what’s running in the cloud, there’s been this very strong trend towards IT/OT convergence. And so things that, a lot of business workloads, increasingly are overlapping with IoT devices. And so it’s very useful to have a platform that can run different business services as well as edge computation. So there’re all kinds of used cases for this, where you might want to combine things in a lot of new and interesting ways.

And one thing in particular, you’re talking about the importance of moving things out of cloud to the edge for the purposes of minimizing bandwidth utilization and latency. And similarly for the real-time computing capabilities, which I believe these are the first core processors that offer that feature, these are both important capabilities in many applications, but one of the things that’s been really growing quickly has been AI applications, which potentially can be very, very data hungry. And I think this is a perfect example of where the computing really needs to happen at the edge. Would you agree with that?

Kenny Chang: Yeah, sure. Absolutely.

Kenton Williston: And that leads me out to another thing I wanted to get your opinion on, is the GPU. So they’re very, very heavily upgraded. And of course, when Intel announced these new parts, the main thing they showed off was how you could play the latest and greatest games on these processors, but maybe not the most relevant thing for industrial healthcare applications. But there is a lot of relevance that it might surprise people, because you can actually use these GPUs to accelerate AI workloads quite a bit. Is that something that you’re seeing as an important way to use these processors?

Kenny Chang: As you mentioned, AI is the mega trend. And right now we can see the Alder Lake-S, they have a big improvements on the GPU performers. As I know they will have 1.94 faster graphic performance, and up to 2.81 faster GPU image classification performance compared to the previous generation. The great benefits for us is we can eliminate the additional GPU card put into our box. That’s good for us, especially in the industry use case, they can reduce lots of maintenance cost on that. We don’t need any simplification of the performance, as well.

Kenton Williston: Yeah, I think that’s all very true. And I think there’s an awful lot of benefit to be had from being able to execute image classification and other AI workloads directly on the CPU. A less complex, easier to maintain lower cost system if you don’t have to add a graphics card. And that I think has always been true. If you can avoid adding more parts, it’s always better, but especially at the moment, graphic cards are very hard to obtain. And of course they’re very expensive when you can get a hold of them. So, I think it’s really nice to have a platform where you can get a tremendous amount of performance out of the GPU right out of the box without having to add any cards. Are there any other benefits to the GPU upgrade? Like I said, probably your customers are not too worried about gaming, but they are still quite upgraded and I’m wondering if you’re seeing any other use cases beyond things like image classification for the GPUs.

Kenny Chang: Another case is about the factory automation. We have the AOI virtually integrated with the AI capability to enhance the capability and the productivity for the defect inspection without a GPU card. They also have the more complex size which can be put into the enclosure. That’s the other key feature for us, to have the same performance, but with the more complex size integrated into our production line.

Kenton Williston: Yeah. That makes sense. And you’re making a very good point that, just the size of the solution by itself can be very important. There are a lot of applications, such as smart city applications, where you might need to squeeze the equipment into an existing space that’s quite limited, or on a manufacturing line where it’s already crowded, any space savings you can get will be very beneficial. That is a very good point. And I’m glad you brought it up. One thing I’m wondering, beyond the hardware attributes, of course, you need to be able to program these things. And I’m wondering from a software perspective, how your customers can best take advantage of all of these new features?

Kenny Chang: Yeah. I think the hybrid architecture, some of the heavy workloads, they need huge powerful processor to deal with that. So, they can assign the task onto the P-cores , what we call the performance core. But some of background, as such as the impact management, they don’t need so much powerful processor to deal with states. They can choose the efficient core to take this job. So this also means they can easily assign: which task, which core. So, I think the major benefits for the software development guys, to respond, to leverage the technology here.

Kenton Williston: And are your customers using things like Intel^® oneAPI to take advantage of this hardware?

Kenny Chang: They are mostly using the OpenVINO^™ for the AI inference task. Intel also has the oneAPI. I think it is good for them to get the API that they want on the one platform. That takes a lot off all their workloads, this platform, I hear from them.

Kenton Williston: That’s really great. And I’m glad you mentioned the OpenVINO architecture. So this is a platform that provides a layer of software abstraction so that you can create and implement different kinds of AI algorithms without having to know all the fine details of the architecture. And it’s very useful when Intel does things like this latest generation Intel core processor, which has a very much faster GPU. You don’t have to worry necessarily about rewriting your code. You just get a performance boost, which is very, very helpful. So, I’m wondering in practice how some of this is playing out. So, I understand you worked with a company called DMS, and you mentioned one of your major lines of business is automated optical inspection. And I understand you did some work with DMS to use the 12th Gen Intel^® Core^™ processors. Can you tell me about some of the challenges that company was facing and how they benefited from using Alder Lake?

Kenny Chang: For the first touch with our customer, they are introduced to the AOI with AI to enhance the accuracy and the efficiency. They really did a good job compared to not introducing the AI task into the AOI. But they also encounter challenges regarding the data transmission. The time between the computer A to computer B, it would take too long to complete the data transition. As you know, the image size is very big, a few megabytes for one image. We saw this challenge and hear the pain points from their viewpoint. Then we think about the Alder Lake-S processor, where it would face their headache a lot. So, we introduced the proposal to them. We would like to build up the workload-consolidation solution. That means we will integrate the computer A and the computer B into one platform. And the most important, we are using the virtualization by the KBN combining the Windows OS and the Linux OS onto the one hardware platform. And we address the data transmission by the shared memory technology. That can make it 100 times faster compared to the previous one.

Kenton Williston: Yeah. And I think this is a really good example. I really appreciate you sharing that with us, Kenny. So, I think many, many factories and other sorts of industrial use cases are in a similar situation right now. AI is the big mega trend right now. It’s being deployed just in all kinds of use cases, but it is very difficult, if you’ve got some existing equipment, to just keep adding some additional equipment to perform AI, because like you said, many times what’s happening with the AI is processing a huge amount of data, whether that’s images or other high bandwidth sensor data. So, sometimes it’s really just the network that is the constraint, even just the local network, never mind going to the cloud, that would prevent you from adding AI to your system. So, the fact that you have very IT-friendly standards based on the platform that can do… It can run windows, it can run Linux, you can have all kinds of different virtualized environments. And you can bring things together on a single platform, making it much easier to add these new capabilities, because you don’t have to have old machine A, new machine B scenario.

It’s like you said, you can just run everything on one machine and support your existing software to a new box, and then start adding all the new things you want. And then in addition to consolidating those workloads, you now have a platform that is very well suited to consolidating other kinds of workloads that you might have in nearby machines or taking things that are currently running into the cloud and bringing them out to the edge.

You have a lot of options there. So, I’m curious, I know that you’re one of the first companies to come to market with a solution for the industrial market based on the 12th Gen Intel^® Core^™ processors, how is that possible and how are you working with Intel to deliver these solutions to market quickly and to not just bring in early solutions to your customers, but really the most advanced kind of solutions?

Kenny Chang: We have a long partnership with Intel for a very, very long time. ASRock Industrial is the leading company for the industrial, mobile, and the system production, especially for the industrial applications. As you mentioned earlier, it is very good for us to get the early sample by the EA program with Intel. We also accept lots of performance updates and the technology updates, such as the tool or architecture to help us deal with the vertical markets, such as the edge insight for industrial and edge control for industrial. They can bring more insights how to address customer needs and how to help our customer, especially for the systems integrator to reduce their development time.

They just focus on what they are good at. They don’t need to waste time to deal with the hardware and software integration. So they just put their application stuff onto the box and that’s done. So they save lots of development and working time from there, and they can get the quick-win solution.

Kenton Williston: So Kenny I’m interested, you’re talking about how your close relationship with Intel and  the early access relationship and access to their roadmap and all these things really help you build application-ready boxes. Can you give me an example of what some of the features might be for some of the solutions you’ve got available now using the latest Intel core processors?

Kenny Chang: Well, that’s good question here. Initially, the feature was for a workload consolidation. More precisely we can say it is the middleware. What we did for customers, just like the case I mentioned before for the AI AOI, we put the virtualization middleware KBN onto our hardware box, and we know how to enable the shared memory and then turn it into API for our customer. So if a customer would like to use this solution, they can just buy our box and we can have such stuff installed in our system, so they can just open the box and put on their software application, then that’s up and running quickly.

Kenton Williston: Yeah, that’s really interesting. So, it sounds like what you’re telling me is ASRock goes beyond simply putting the hard hardware together and sending the hardware to your customers, but you actually offer a certain level of services to provide some appropriate set up and middleware and these things. So that when the system arrives at the customer, it’s ready for them to start putting their software on it. They don’t have to think about those things. Do I have the right idea there?

Kenny Chang: I think it’s just a very beginning stage. It’s optional items. If customer needs this, we can do such a service for them. I think the right now we will take some time to educate the institution, more ideal to our customer before they adopt the solution.

Kenton Williston: Got it. That makes sense. I wanted to, speaking of education, touch on something we haven’t talked about yet. One of the other new features of the platform or some new hardware security features, do you think these will be important for your customers as well?

Kenny Chang: Well, yes, it’s very, very important. Cybersecurity in tech is the hot topic all over the world. In most cases it’s happening in the IT, but right now, when we introduce the industrial IoT into the industrial automation, there are lots of OT devices. They are very vulnerable.

Let me bring one case I have. The case is about 5G smart pole, smart city. Smart pole is integrated with a lot of devices, not only for the lightning for the streets, but also it has the sensor for the air condition quality. Also, we have the smart camera to monitor the traffic overall in the city. We provide our Alder Lake-S platform solution into the smart pole, which working as the edge server.

All the camera data will go into the edge server for the image classification or any of the dual-based checking by the edge server. But all this data is very sensitive. The benefits to adapt the Intel processor is they are integrated with the software guards we call SG C, as well as they have the PTT technology. So they can make the data be secure in the hardware method. And they also have very good leverage by this from the system integrator to put the software on that and to ensure the security be in place.

Kenton Williston: That all makes total sense. You’re raising some good points about the Intel software guard, or SWG, and the PGP features that are in there as well. It’s very important, like you said, many more systems are coming under attack, and earlier I was mentioning how it’s a good thing that IT and OT are converging and the systems that used to have very specialized code are now much more often becoming, by necessity, more IT-friendly systems that run very familiar operating systems.

This is a good thing because it helps innovation move forward more quickly, but also makes these systems more open to attack than they used to be. I like this example you gave of a smart pole where you might have lighting and cameras and other sensors, and it’s very useful to have real-time visibility into what’s happening in the city, whether it’s air pollution or traffic levels or whatever.

But of course, especially with cameras, there’s always a risk that people could get access to video feeds that they really shouldn’t have access to. And so it is very important to keep in mind security, and of course just the very fact you’re talking about something that’s connected back to government systems.

There can be all sorts of very subtle ways of… once you’ve gotten into a system getting back into very sensitive data, and there’s been cases of people, for example, showing how you can access, say a printer, for example, and then just a couple of jumps and you’re into a very sensitive database. Yes. So, the security is very, very important. So I’m glad we had a chance to talk about that. We’ve covered a lot of different topics. I’m wondering if there is anything that I didn’t ask you about that you would like to add.

Kenny Chang: I just want to add that ASRock Industrial is not only for the hardware provider, we also think about and are working with the Intel verticals to help customers to get the better solution for them. That’s our goal for our end customer. And we also help, by this, co-create to make the world  much better than ever.

Kenton Williston: I love that. That’s great. Well, Kenny, I want to thank you again for joining us today. I really appreciate your time.

Kenny Chang: Yeah. Thank you.

Kenton Williston: And thanks to our listeners for joining us. To keep up with the latest from ASRock Industrial, follow them on LinkedIn at ASRock-Industrial, that’s ASRock-Industrial.

If you enjoyed listening, please support us by subscribing and rating us on your favorite podcast app. This has been the IoT Chat. We’ll be back next time, with more ideas from industry leaders at the forefront of IoT design.

The preceding transcript is provided to ensure accessibility and is intended to accurately capture an informal conversation. The transcript may contain improper uses of trademarked terms and as such should not be used for any other purposes. For more information, please see the Intel^® trademark information.