In The Arena by TechArena - Accelerating Simulated Design to Fuel Innovative 5G and Edge Use Cases with Ansys

Episode Date: March 26, 2024

TechArena host Allyson Klein chats with Ansys Chief Technologist Christophe Bianchi about his company’s mission to deliver design simulation across industries, how the communications arena represents an opportunity from silicon to systems of systems, and how AI is accelerating capabilities in Ansys solutions.

Welcome to the Tech Arena, featuring authentic discussions between tech's leading innovators and our host, Alison Klein. Now let's step into the arena. Welcome to the Tech Arena. My name is Alison Klein, and today we're coming from Mobile World Congress in Barcelona, and I'm so delighted to be joined by Dr. Christophe Bianchi, Chief Technologist for Ansys. Welcome to the program. Thank you. Thank you. So why don't we just start? Ansys has never been on the Tech Arena before. Why don't you just start with an introduction

of the company and your role? Yes. And actually, also, the first time we were at Moldova Congress, so we are here to cover that. Ansys is a company that's been in business for more than 50 years. And we specialize in just one thing. We do engineering simulation. What we do is convert the law of physics into a program to help designers in all industries create their products. That's what we've been doing and it's covering multiple aspects of physics from aerodynamics

to electromagnetic wave propagation, which is very specific to what we see here at Mobile World Congress, but we cater to the automotive industry, we cater to the aerospace and defense, and of course to what we call the high-tech industry, which is semiconductor and intelligent connectivity. My role in the company, I'm chief technologist, I report to our CTO. The way we are organized is by industry, so my focus is on high-tech. That doesn't mean that everything else is low-tech, but it's really where most of the electronic intelligence is built into products and end products. I'm based in the south of France,

where I cover our customers worldwide. And the role of the CTO office is to work ahead of the curve. So we are working with engineering department, R&D departments, and our customers, which means that the product we're helping them build will hit the market in two to three years. The role of the CTO office is anticipate the next wave of demand. So we're working on the two or three-year horizon of our solutions

to probably hit the market in six, seven years. I'm very familiar with Ansys in your role in high performance computing and being one of the largest workloads in terms of compute performance with my background in semiconductors. But your name was new to me from a standpoint of Mobile World Congress. And I guess one place that I want to ask is, how do you approach simulation and design acceleration for the communication arena? And what are the opportunities there?

So, and that's probably known only to the engineers today. We're a well-kept secret. Not by design, but that's the way we're written. Too shy, I don't know. When it comes to electromagnetic simulation, radio frequency propagation, we've been the gold standard for many years. We have solutions,

brand name is called HFSS, every designer building, antennas, base stations, communication systems in all different wavelengths and bands is using LNC's tool. If you look at the companies present at Mobile World Congress, all the companies building equipments are user-friendly.

But until now, until very recently, we only focused on the equipment, the components level. We spent a lot of effort in developing solutions that address the entire communication system from what would go the nanometer to the kilometer. being able to model antennas, phase array antenna, base stations, tower systems, radomes, but also putting this into a silly media, modeling the communication from tower to lower, to subscriber, to satellites. And this is what we're showcasing to our customers today. So moving really from the component to the system, to the system of systems.

Until very recently, it was inconceivable and unfeasible to have the level of accuracy that we provide in our solution, which is true to physics, electromagnetic simulation, at the system level. And it took us a few acquisitions, years of RMD, dedicated hardware working on GPUs to run our software, to be able to deliver what we're presenting this year at the World World Congress, which is called the entire channel modeler, the ability to run real-time simulation of full-digit between of a complete

5G operating system, 5G network, modeling all the channels, all the signals at the same frame per second, which then runs actually to the point that we can't even connect this API to our software to hardware testers or 5G testers. And this is unlocking a lot of new use cases for simulation, for digital trends. And it's becoming essential because what we see in the recent deprivation of IG and the upcoming CG is the demand for more customization, more use case, more ability to do software

defined everything, whether it's software defined satellite, like the conversation we had this morning, or software defined our communication system, software defined vehicle, all these are the components of recordability while the product is operating. And for that to do that in a flying vehicle, all these are components of recordability while the product is operating. And for that to do that in a predictable fashion and obtain the benefits

of seeing reliability, safety, and performance, concept-altering simulation, evidence, digital training is essential. That's what we've introduced in our knowledge here in the lower part. Now, when you were talking, one of the things that I was thinking about is the conversations from operators on the expense of truck rolls. And the expense, especially when you consider the broad span of networks and what you were just describing in terms of being able to see that entire picture. Can you talk a little bit about how simulation has impacted your customer base and where the gains that they see? Is it time? Is it money?

Is it ultimate solution performance? Or is it all of it? It's a combination of different KPIs and different metrics, but if I look at the component level, when we interface with our customers with home, simple problem like developing an antenna, through simulation you can reduce the number of prototype disciples and add the right first line results, that's the only time it's also opportunity cost. So we see reduction in the order of development cycle when we implement the full suite of simulation. So basically creating the entire antenna in the virtual world

while it's lying in before you start to build the first physical evidence. And that goes with not just the antenna system, but the complete other base station, the complete network infrastructure. So simulation is a way to reduce trust by reducing another physical prototypes. It's also reducing, of course, the design time in orders of magnitude. But because of that, because of being in the virtual world,

you actually unlock much more creativity from a design standpoint. And a lot of designers will explore much broader design spaces. When you apply that to network planning, for instance, you have the ability to simulate the performance of your network in a given city. You upload a very precise

city of Detroit, for example, we're running here with five centimeter level accuracy. You can have all the KPI analysis and all the performance of your network and decide where to put antennas. So optimizing the setting of your network is also a cost for the operator because it's going to have the right performance

and the right fit between what the system provider is going to bring to the operator. And we're at Mobile World Congress, so I've got to ask, you know, everybody in the hallways is talking about broad proliferation of 5G. Are we seeing a downturn in 5G infrastructure deployments, the upcoming 6G specs, and are we seeing a turn to starting to develop 6G solutions? How do you see that evolving, and why do you feel like there's this interplay between 5G and 6G right now? So what we see, and again, it's from the very, very small view of the world, which is working in ways that are very important.

First of all, 5G is important, but not really utilized. If you look at the average usage of the mini-interface section of the intersection of 5G network, it's very small. It's around 10%. And these are use cases that are generally not implemented, but these are eye-bending use cases. So working on these use cases requires

also continuous development and improvement of the install hardware. When these new use models will be deployed, the image slicing, everything that has to do with providing more value and more customization on a subscriber level of the quality of service, that continues to require a lot of engineering effort and development. It's not just in the software side. It's really at the interface of the

software running on this network system and the hardware. This is where exploring the use case, exploring how far we can push 5G is a place where simulation is really critical. 6G, it's

a little bit more than research now. We start to see a system that we are testing. We're moving in a completely different world because frequencies are very different, much higher. But also the reconfigurability, the ability to dynamically reconfigure the constellation to optimize the traffic on these frequencies is extremely complicated. And again, having an accurate way of modeling that ahead of time is important. 6G, to my knowledge, is a slightly different use model than 5G.

It doesn't address the same needs. It's definitely not for the general deployment just the way 5G is. Even in 5G, working down the aisles of the World War Congress, which I

didn't have time to do as much as I wanted. But it's really, you hear

a lot about private network versus public network. So applications that are

really dedicated to one particular industry called zero that we talked about.

We've seen applications in the healthcare sector where you really want to control the quality of service. We're going to continue to explore more use cases of 5G, which I think for the next decade is still going to be the main, dominant thing, and 6G, we're going to pick up additional use cases. No interview would be complete in 2024 without a mention of AI. And I know Ansys has put the technology into practice in a lot of your solutions.

How is it forming your decision-making around technology choices? And where can you point to examples of how AI has really transformed what you're able to do for your customers. So besides the very popular large language model, which has been a function, which is basically a replacement of the traditional user's manual, where you're just querying and using natural language, that every company does that. For us, AI is way more than this. There are basically, there are basically two aspects of AI in our

offering and the way we interact with our customers. The Eden part, which is using AI as a way to make our product provide better results and faster results. Using machine learning, coding techniques to

improve our generic solvers and our specific solvers. That's been practiced in our engineering department for the past decade. We have today close to 95% of our software that has, to some extent, a portion of AI built in. The second part is how do we use AI to enable more design productivity to our customers? And I usually take a very, very simple example. Simulation just is a way to virtualize your design environment. It tells you what a given design will look like. And a good engineer can, if you give a design with three dimensions,

three type of parameters, you can option. So you're in three dimensions. We can visualize where the optimum solution would be. Good engineers can probably think in six dimensions. Nobel Prize probably in seven or nine. But today the parameters that dictate what a 5G network looks like are hundreds. Right. There's absolutely no way we can visualize or interpret or come up with the optimal configuration for a given product.

And naturally, without machine learning and AI, the way to do that is you simplify the problem by just restricting your design space by fixing some of the parameters and exploring just as you are. What AI does, it unlocks all these dimensions. So with AI-powered solutions, and we have multiple flows that give access to these AI technologies to our customers,

we have the ability to explore all potential options for new or for improvement of a design. There's also a second element which is essential is AI is, when we talk about AI in our space, it's a lot about machine learning. In machine learning, there's

learning in it, so it needs a lot of data. But in our space, data is simulation data. And our customers, they're not new to electromagnetic simulation. They've done this for decades. So they have seen it on their shelves.

You have assets that are catching dust, which are simulation results of previous generations. This is a goldmine for machine learning, and what we've done is we introduced methodology and solution that actually turn these dormant assets into IP, which is the customer IP.

There's several of them. That we really, that is an input to provide response models that will guide the designers into improving or creating new solutions. The data is captive to the customer. We don't look at the data. That's the customer data. That's the number one principle in AI is customer's data is by the customer.

But we provide solution in metal so they can have these dominoes that's into value. Very nice. Now, we're at MWC. I've got to ask you, you've been in a lot of industry conversations. It sounds like your team never stops with industry conversations. What are the highlights from others in the industry about what's happening at MWC this year? And what are you most excited about? Oh, so, I'm excited, well, the excitement comes from

the realization that there's more use case than just the regular use of our cell phone. We're working

in many industries. One of the industries that is evolving extremely fast is automotive. Automotive, if you look at

level four autonomy, we all know that it's not going to happen without V2X and V2V. That requires infrastructure. Cartel drive itself autonomously in a city,

there has to be five-year work. How do we bring these two worlds together? Because it used to be that in engineering, whether it's a semiconductor, it may not matter that you know very well, or other. In the last decade, you would give a specification to the designer,

I would design the product to spec, and then this job would be finished. Today, the spec is the system. The system is multiple car driving together, obviously, that's the spec. How do we help facilitate this conversation is what everyone is trying to solve here. And we see a very interesting solution in transportation with drones, with autonomous driving.

Industrial IoT is now catching up. And that requires a lot of simulations. We have a job for generations. A lot of complicated simulations. That's fantastic. It seems like AI and some of those new techniques are coming at exactly the right time. Yes, for us and for our customers. We see very impressive AI use cases in the private section of the demo of some of our top customers.

I can't name them, but I'm still trying to figure out and understand exactly the magic they were presenting to me. So I would give a lot of technology if this is the place to be. That's fantastic. One final question for you, Christophe. Where can folks find out more about Ansys and the solutions you talked about today? So, of course, on our website,

there's a lot of information. They can reach out to us through all the social network, your coordinates. I'm happy to talk about the future of technology in all industries, and particularly in IT and telecommunication. Thank you so much for spending 30 minutes of your time with me. I know your time is precious this week. Thank you. Thank you so much. Welcome to the Tech Arena, featuring authentic discussions between tech's leading innovators and our host, Alison Klein. Now, let's step into the arena.