 theCUBE's live coverage is made possible by funding from Dell Technologies, creating technologies that drive human progress. Hey everyone, welcome back. Good evening from Barcelona, Spain. It's theCUBE, the leader in live tech coverage, as you well know, Lisa Martin and Dave Nicholson. Day two of our coverage of MWC23. Dave, we've been talking about sexy stuff all day. It's about to get, we're bringing sexy back. It's about to get hot. It's about to get hot. We've had two guests with us, two senior consultants from the product planning, networking, and emerging server solutions group at Dell, Heather Rayhill and James Bryan. Welcome guys. Thanks for having us. Really appreciate it. Dave, you're bringing sexy back. I know, we are, we are, we wanted to bring, yes. We've been talking about this all day. It's here. Yes. Talk to us about why this is so innovative. So, actually we wanted to bring this, getting a lot of attention here. As a matter of fact, we even have a lot of our competition taking pictures of it. Why is it so innovative? So, one of the things that we've done here is we've taken a lot of insights and feedback from our customers that are looking at 5G deployments and looking at how do they basically bring commercial off the shelf to a very proprietary industry. So what we've done is we've built a very flexible and scalable form factor in the XR8000. And so this is actually a product that we've purposely built for the telecommunication space. Specifically, it can be deployed for surfing a virtual DU or a DUC at a cell site for distributed RAM or it can be put in a local data center but outside a main data center to support centralized RAM. We'll get into it, which is where the early excitement gets, is, it's flood based in its design. And so, because of that, it enables us to provide both functionality for telecommunications, could be network, could be enterprise edge, as well as being designed to be configured to whatever that workload is and be cost-optimized for whatever that workload is. All right, you're killing us. Let's see. Show it, show it to us. This is where I have to hand it off to my colleague Heather. But what I really want to show you here is the flexibility that we have and the scalability. So right here, what I'm going to show you first is a 1U sled. So I'll set that out here and I'll let Heather tell us all about it. Yeah, so X8000, let's talk about flexibility first. So the chassis is a 2U chassis with a hot swap shared power supply on the right. Within it, there are two form factors for the sleds. What James brought out here, this is the 1U form factor. Each sled features one node or one CPU for sled. So we're calling the 1U the highest density sled, right? So you can have up to four 1 node 1U sleds in the chassis. The other form factor is a 2U sled on the right here and that's just really building on top of the 1U sled that adds two PCIe slots on top. So this is really our general purpose sled. You can have up to two of these sleds within the chassis. So what's really cool about the flexibility is you can plug and play with these. So you could have two 1Us, two 2Us or a mix and match of each of those. Talk about the catalyst to build this for telco and some of the emerging trends that you guys have seen and said this needs to be purpose built for the telco. There's so much challenge and complexity there, they need this. Let me take this. So actually that's a great question by the way. It turns out that the market's growing, it's nascent right now. Different telecommunication providers have different needs, their workloads are different. So they're looking for a form factor like this that when we say flexible, they need to be able to configure it for theirs. They don't all configure the same way. And so they're looking for something that they can configure to their needs, but they also don't want to pay for things that they don't need. And so that's what led to the creation of this device, the way we've created. How is it specific for edge use cases? So if we think of the edge, it's emerging, it's burgeoning. What makes this so specific to edge use cases? Yeah, let's talk about some of the ruggedized features of the product. So first of all, it is short depth, so only 430 millimeters. And this is designed for extreme temperatures, really for any environment. So the normal temperatures of operating are negative five to 55, but we've also developed an enhanced heat sink to get us even beyond that. That's Celsius. Celsius, right, thank you. So this will get us all the way down to negative 20, boot in operating, all the way up to 65C. So this is one of the most extreme temperature edge offerings we've seen on the market so far. And so this is all outside the data center. So not your typical data center server. So not only are we getting those capabilities, but half the size when you look at a typical data center server. So these can go into a place where there's a rack, maybe, but definitely doesn't have to be raised floor. It is still size cabinet. It's power, yeah, okay. Yeah, and we also have AC and DC power options that can be changed over time as well. So what can you pack into that one, one use sled in terms of CPU cores and memory, just as an example? Yeah, great, so each of the sleds will support the fourth generation of Intel Sapphire Rapids up to 32 core. They'll also be supporting their new VRan Boost SKUs. And the benefit of those is it has an integrated feck accelerator within the CPU. Traditionally to get feck acceleration, you would need a PCIe card that would take up one of the slots here. Now with it integrated you're freeing up a PCIe slot and there's also a power savings involved with that as well. So talk about the involvement of the telco customer here and then design. I know Dell is very tight with its customers. I imagine there was a lot of communications and collaboration with customers to deliver this. Interesting question. So it turns out that early on, we had had some initial insight, but it was actually through deep engagement with our customers that we actually redesigned the form factor to what you see here today. So we actually spent significant amount of time with various telecommunication customers from around the world and they had a very strong influence in this form factor. Even to the point, like I mentioned, we ended up redesigning. Do you have a sense for how many of these or in what kinds of configurations would you deploy in the typical BBU? So if we're thinking about radio access network, literally, tower, transmitter, receiver, somewhere down there in a cabinet, you have one of these, you have multiple units. I know the answer is it depends. You are right. But if someone tells you, well, we have 20 cellular sites and we need, we're moving to an open model and we need the horsepower to do what we want to do. I'm trying to gauge what one of these, what does that mean? Or is it more like four of these? So it depends. It depends. You're absolutely right. However, we can go right there. So if you look in the to you, we have three PCIe slots, as Heather mentioned. And so let's say you have a typical cell site. We could be able to support a cell site that could have three radios in the configuration here. It could have, multiply by three, right? It could have up to 18 radios. And we could actually support that. We could support multiple form factors or multiple deployments at a particular cell site. It really then, to your point, it does depend. And that's one of the reasons that we've designed it the way we have. For example, if a customer says their initial deployment, they only need one compute node because maybe they're only going to have, two or three carriers. So then there you've got maybe six or eight or nine radios. Well, then you've put in a single node, but then they may want to scale over time. Well, then you actually have a chassis. They just come in and they put in a new chassis. The other beauty of that is, is that maybe they wait, but then they want to do new technology. They don't even have to buy a whole new server. They can update to the newest technology, same chassis, put that in, connect to the radios and keep going. But in this chassis, is it fair to say that most people will be shocked by how much traffic can go through something like this in the sense that if a tower is servicing and number of conversations and data streams, going through something like this, I mean, somehow it blows my mind to take a thousands of people accessing something and having them all route through something like this. It will depend on what they're doing with that data. So you've probably talked a lot about a type of radios. Are we going to be massive MIMO or what type of radios? Is it going to be a mix of 4G or 5G? So it'll really depend on that type of radio and then where this is located. Is it in a dense urban environment or is it in a rural type of environment at that cell site shelter, but out in a suburban area? So it will depend, but then that's the beauty of this, is then I get the right CPU. I get the right number of adding cars to connect to the right radios. I purchase what I need. I may scale to that. I may be in a growing part of the city, like where I'm from or in San Diego, where Heather's from, where she's in a new suburban and they put out a new tower and the community grows rapidly. Well, then they may put out one and then you may add another one and I can connect to more radios, more carriers. So it really just comes down to the type and what you're trying to put through that. It could end at a stadium where I may have a lot of people. I may have video streaming and other things. Not only could I be a network connectivity, but I could do other functions like Mac, Multiaxis X on point that you've heard about, talked about here. So I could have a GPU processing information on one side. I could be network on the other side. I do, I do. Go for it. No, no, no, I'm sorry. I'm sorry, I'm doing a hog all of the time. What about expansion beyond the chassis? Is there a scenario where you might load this chassis up with four of those nodes, but then because you need some type of external connectivity, you go to another chassis that has maybe some of these sleds or are these self-contained and independent of one another? They are all independent. Okay. And then we've done that for a reason. So one of the things that was clear from the customers again and again and again was cost, right? Total cost of ownership. So not only how much does this cost when I buy it from you what is it gonna take to power and run it? And so basically we've designed that with that in mind. So we've separated the compute and isolated the compute from the chassis from the power. So I can only deal with this. And the other thing is, is it's a sophisticated piece of equipment that people that would go out and service it are not used to. So they can just come out, pull it out without even bringing the system down if they've got multiple nodes, they don't have to pull out a whole chassis or whole server. Put one in, connect it back up while the system is still running. If a power supply goes out they can come and pull it out. We've got one. It's designed with a power infrastructure that if I lose one power supply I'm not losing the whole system. So it's really that serviceability total cost of ownership at the edge which led us to do this as a configurable chassis. I was just going to ask you about TCO reduction but another thing that I'm curious about is there seems to be like a sustainability angle here. Is that something that you guys talk with customers about in terms of reducing foot print and being able to pack more in with less reducing TCO reducing storage power consumption, that sort of thing? Go ahead. We need to take that one as well. So yeah, so it comes to be various by the customer but it does come up. And matter of fact, in that vein similar to this from a chassis perspective is I don't, especially now with the technology changing so fast and customers still trying to figure out well, is this how we're really going to deploy it? You basically can configure and so maybe that doesn't work, they reconfigure it. Or as I mentioned earlier, I purchased a single sled today and I purchased a chassis. Well then the next generation comes. I don't have to purchase a new chassis. I don't have to purchase a new power supply. So we're trying to address those sustainability issues as we go again back to the whole TCO. So they're kind of related to some extent. Right, right, definitely. We hear a lot from customers in every industry about ESG and it's an important initiative. So Dale being able to help facilitate that for customers I'm sure is part of what gives you that competitive advantage. But you talked about James that and we talked about it in an earlier segment that competitors are coming by sniffing around your booth, what's going on. Talk about from both of your lenses the competitive advantage that you think this gives Dell in telco. Heather, we'll start with you. Yeah, I think the first one which we've really been hitting home with is the flexibility for scalability. This is really designed for any workload from AI and inferencing on a factory floor all the way to the cell site. I don't know another server that could say that all in one box. And the second thing is really all of the TCO savings that will happen immediately at the point of sale and also throughout the life cycle of this product that is designed to have an extremely long lifetime compared to a traditional server. Yeah, I'll get a little geeky with you on that one. Heather mentioned that we'll be able to take this eventually to 65C operating condition. So we've even designed some of the thermal solutions enabling us to go there will also help us become more power efficient. So again, back to the flexibility even on how we cool it so that it enables us to do that. So do you expect, you just mentioned maybe if I heard you correctly the idea that this might have a longer user usable life than the average kind of refreshed cycle we see in general IT. What I mean, how often are they replacing equipment now in kind of legacy network environments? I believe the traditional life side of a server is what, three to five years? Three to five years traditionally and with the sled based design like James said we'll be designing new sleds, you know every year, two years that can just be plugged in and swapped out. So the chassis is really designed to live much longer than just three to five years. We're having customers ask anywhere from seven to when it dies. So substantial increase in the life cycle as we move out because as you probably know well, right? The further I get out on the edge, the more costly it is. And I don't want to change it if I don't have to. And so something has to justify me changing it. And so we're trying to build to support that both that longevity but then with that longevity things change. I mean seven years is a long time in technology. So we need to be there for those customers that are ready for that change or something changed and they want to still be able to adopt that without having to change a lot of their infrastructure. So customers are going to want to get their hands on this obviously we know, we can tell by your excitement is this GA now, where is it GA and where can folks go to learn more? Yeah, so we're here at Mobile World Congress in our booth. We've got a few featured here and other booths throughout the venue. But if you're not here at Mobile World Congress this will be launched live on the market at the end of May for Dell. Awesome and what geographies? Worldwide. Worldwide. Put your hands on the XR8000 worldwide in just a couple of months. Guys thank you for the show and tell talking to us about really why you're designing this for the telco edge, the importance there. What is going to enable operators to achieve? We appreciate your time and your insights and your show and tell. Thanks. Thank you. For our guests and for Dave Nicholson, I'm Lisa Martin. You're watching theCUBE live from Spain in mobile MWC23. Be back with our show day two wrap with Dave Vellante and some guests in just a minute.