Richard Barrett, Hewlett Packard Enterprise (00:06):
Hey, good afternoon guys. I hope you are all having a good show. I'd also like to introduce Jose Arias, my colleague, and we're here to talk about digital twin. So just so we know, for those who don't know what a digital twin actually is, a digital twin is a digital representation of a physical object or a theoretical object. So this could be anything. We look at this as pretty pictures, but imagine if these objects are all defined as individual objects. The lemon, the bread, all has texture, all has the ability to rot like a real, real vegetable, real bit of fruit. You can actually get it to move. You can do things that are quite interesting with these devices and these objects across the board. You can do it at massive scale with fire. You can actually use physics to look at how objects act in the water.

(00:55):
You can do this at massive scale. You can do it at water tanks and you can do this in warehouses. Data centers, we can look at this in terms of how millions of objects actually react with each other, how they impact each other, how they fall over looking at safety, looking at obviously things that go wrong. How can we avoid those things going wrong? How we can simulate orbits at massive scale to make sure that actually when we go through the idea of prototyping these things, we can do this multiple times. They have real physics in it, but also the fact that we look at these things can actually be used as showcases for the products before they're built. But you can make them at such fidelity that there are actually beautiful things to look at. You can also integrate things where you're looking at weather patterns.

(01:46):
Digital twins are based on data sets that you can take and data sources that you pull in together and effectively you can make that into a platform that does everything from the building blocks all the way up through robotics, how robotics work at massive scale, how all these platforms work together. Robotics, how will this dog actually jump? You can look at digital twins for sporting events, sporting prowess. What's the best thing you can do? What about looking at traffic analysis? How does the traffic flow? All of these things you can do before you ever build everything before you put a spade in the ground. Looking at telecoms, looking at factory capability. And of course the other beauty of this is looking at it, it's beautiful. You can look at this with VR headsets, you can interact with it with natural language. You can do a whole host of fantastic things.

(02:43):
So the problem is, as I mentioned, you start with a digital twin. So the virtual representation of your device or your object, the problem is you've got data sources all over the place. You need to establish which of these data sources are going to give you the right data. The old adage of garbage in garbage out is really true here. So effectively if you get the data wrong, the digital twin will be wrong and HP can help you on this journey. Going through building the data repository, moving this into actually the creation, the digital twin, and then more importantly how you actually interact with that. Be it on a 2D screen, how you act with it in terms of virtual reality, how you interact with it in terms of multiple people. It can change the user experience, it can change how you bring products to life.

(03:33):
It can change how you do return of investment based on lifecycle management of a product. And there's lots and lots of use cases. Anything from networking where you can use this. There's a great demo on, I'll stand here around Ericsson who use this to look at how signals propagate. They built a digital twin of the local office environments and external environments to see how data paths and network paths propagated around buildings. So signal so they could use that to assess what was the right location to put transmitters and receivers. So all that was modeled long before they actually did the implementation of run, et cetera, so they could see how this was going to interact with the real world and optimize where they located things. Same with assembly line. We have people like Mercedes-Benz, who have spent a lot of time building the digital twin of the production line so that when they actually put a spade in the ground, they know that production line is optimized.

(04:32):
They know it works. They know where things like predictive maintenance and needs to take place because they've actually used it in a real world. Those objects have got real world properties. So when the two objects touched each other, they understand how they were. So they know the life cycle of certain parts may be two years. So at 20 months they replace that part. So there's no one scheduled downtime. There's a huge number. And of course we saw something around warehousing. Amazon used the similar product stack to develop an entire warehouse before they actually built it to make sure it all worked. We do this at the moment with products like that. Jose's going to introduce you in a second around Nvidia Omniverse, VMware Horizon, and a product stack from HPE, which solves a lot of the problems I'm going to talk about, which is this all sounds really good.

(05:23):
This all sounds like a great thing we can use to build multiple products. The look and see what works to simulate how things will act in the real world without spending all that letter. But the problem is at the moment, VMware platforms are usually scattered around the world with various different applications trying to do either virtual reality or digital twins. The teams are disconnected, lack of trust, many vendors, there's probably 60 to a hundred different tools you can use to create a digital twin. And actually getting them all to operate together is really, so compatibility and data management is really important. We all know an example where a large airplane manufacturer had one design function in one country, half of it in another country using two different bits of software. When they came to physically build the aircraft, the bits didn't fit because they'd used incompatible software, siloed data. So getting all this right is really important and the incompati drives inefficiencies. So what I'm going to do is we've solved this problem and at this point I'm going to hand over to Jose to go through what we've done.

Jose Arias, Hewlett Packard Enterprise (06:31):
When you're building a digital twin, obviously you need hardware, software services, and as HP, we want to keep all the complexity regarding the platform here with us. So you can focus in the digital twin, twin Green Digital Twins basically is a complete solution with hardware. As you can see here, we have the compute and we have the storage needed to run digital twins over this storage. We have three layers of software. We have Bmware Horizon and we have Omniverse over this. So with this complete hardware and software solution, we are wrapping everything under Grid Lake. So we are going to provide this as a service for you. What is the meaning of that? That you are going to have a cloud experience. If you are using more hardware and more services, you are going to pay as per use, you are going to pay as your growth. So in this view, we are going to be focused as HPE in providing you the platform, software services and the hardware. And you are going to have the focus building the detailed twins for your own business

Richard Barrett, Hewlett Packard Enterprise (07:47):
And it's worth out, this is entirely certified stack from Nvidia. So it comes out of the box ready to run. All the licenses are included, the platform's included, storage is included. It's run with Nvidia, A 40 GPUs, I think it is in the current flies, but we actually pay, this can scale. It's flexible. So if you decide you want less users, we can scale it back down. You reduce your costs. If you want more users, we can expand the system and expand the platform so you can cope with all those additional users. And I think at this stage, what I'd like to do is say thank you very much. If you have any questions, please let us know guys.