 We're back at super computing 22 and Dallas winding down the final day here Big show floor behind me lots of excitement out there. Wouldn't you say Dave? Oh, it's crazy I mean any anytime you have NASA presentations going on and and steam punk iterations of cooling systems that you know, it's it's a great hundreds of trade shows I don't think I've ever seen NASA exhibiting at one like they are here Dave Nicholson my co-host I'm Paul Gellin and which with us is Satish. I air he is the vice president of emerging services at Dell Technologies and Satish. Thanks for joining us on the Cube. Thank you. Paul. What are emerging services? emerging services are actually The growth areas for Dell So it's telecom. It's cloud. It's edge So we especially focus on all the growth vectors for for the company and And one of the key areas that comes under your jurisdiction is called apex now I'm sure there are people who don't know what apex is. Can you just give us a quick definition? Absolutely. So apex is actually Dell's for a into cloud and I manage the apex services business So this is our way of actually bringing cloud experience to our customers on prem and in color But it's not a cloud. I mean you don't you don't have a Dell cloud, right? It's it's infrastructure as a service. It's infrastructure and platform and solutions as a service Yes, we don't have our own equivalent of a public cloud, but we want to you know, this is a multi-cloud world So technically customers want to consume where they want to consume So this is Dell's way of actually, you know supporting a multi-cloud strategy for our customers You you mentioned something just ahead of us going on air a great way to describe apex To contrast apex with capex. There's no C. There's no cash up front necessary Yeah, I thought that was great explain that explain that a little more I mean, you know one of the main things about cloud is the consumption model, right? So customers would like to pay for what they consume. They would like to pay in a subscription They would like to not prepay capex ahead of time. They want that economic option, right? So I think that's one of the key tenets for anything in cloud So I think it's important for us to recognize that I think apex is basically a way by which customers pay for what they consume Right, so that's absolutely a key tenet for how we want to design it. So it's absolutely right And among those services are high-performance computing services now I was not familiar with that as an offering in the apex line. What constitutes a high-performance computing apex service? Yeah, I mean, you know, I mean this conference is great. Like you said, you know There's so many HPC and high-performance computing folks here But one of the things is, you know, fundamentally if you look at that performance computing ecosystem, it is quite complex, right? And when you call it as an apex HPC or apex for Frank offer It brings a lot of the cloud economics and cloud, you know Experience to the HPC offer. So fundamentally, it's about our ability for customers to pay for what they consume It's where Dell takes a lot of the day-to-day management of the infrastructure on our own So that customers don't need to do the grunge work of managing it And they can really focus on the actual workload which actually they run on this HPC ecosystem So it is it is high-performance computing offer, but instead of them buying the infrastructure Running all of that by themself We make it super easy for customers to consume and manage it across, you know Proven designs which Dell always implements across these verticals So what makes it a high-performance computing offering as opposed to to a rack of PowerEdge servers? What do you add in to make it HPC? Oh, that's a great question. So, I mean, you know So this is a platform, right? So we are not just selling infrastructure by the drink So we actually are fundamentally based on, you know, we launch two validated designs One for life sciences one for manufacturing So we actually know how these peace parts work together how they actually are validated design tested solution And we also it's a platform. So we actually integrate the softwares on the top So it's just not the infrastructure. So we actually integrate a cluster manager We integrate a job scheduler. We integrate a container orchestration layer So a lot of these things customers have to do it by themself, right? If they buy the infrastructure So by basically we are actually giving a platform or an ecosystem for a customer to run their workloads So make it easy for them to actually consume those Now is this is this available on-premises for a customer? Yeah, so we we make it available customers both ways So we make it available on-prem for customers who want it We know kind of they want to take that take that economics We also make it available in a colo environment if the customers want to actually, you know Extend colo as that on-prem environment. So we do both What are the requirements for a customer before you roll that equipment in? How do they sort of have to set the groundwork for? Well, I think you know fundamentally it starts off with what the actual use case is right? So if you really look at you know the two validated designs we talked about you know one for you know health care Life sciences and one other one for manufacturing They do have fundamentally different requirements in terms of what you need from those infrastructure systems So you know the customers initially figure out okay I'll do they actually require something which is going to require a lot of memory intensive loads or do they actually require something Which has got a lot of compute power so you know it all depends on what they would require in terms of the workloads to be and then We do have t-shirt sizing so we do have small medium large we have you know multiple infrastructure options CPU core options Sometimes the customer would also want to say you know what as long as the regular CPUs I also want some GPU power on top of that so those are determinations So typically a customer makes as part of the ecosystem right and so those are things which would they would talk to us About to say okay, what is my best option in terms of you know kind of workloads? I want to run and then they can make a determination in terms of how they would actually go So this this is probably a particularly interesting time to be looking at something like HPC via Apex with With this season of rolling thunder from various partners that you have you know We're all expecting that Intel is going to be rolling out new CPU sets From a PowerEdge perspective you have your 16th generation of PowerEdge servers coming out PCIe Gen 5 and all of the components from partners like NVIDIA and Broadcom etc. plugging into them What what is that? What does that look like from your from your perch in terms of talking to customers who may be maybe they're Doing things traditionally and they're likely to be not not not 15 G not generation 15 servers Yeah, but probably more like 14. Yeah, you're offering a pretty huge uplift. Yep. What what are those conversations look like? Well, I mean so talking about partners, right? I mean of course Dell, you know, we don't bring any solutions to the market without really working with all our partners Whether that's at the infrastructure level like you talked about you know Intel AMD Broadcom, right? All the chip vendors all the way to software layer, right? So we have cluster managers. We have communities orchestrators So we usually what we do is we bring the best in class, whether it's a software player or a hardware player, right? And we bring it together as a solution So we do give our customers a choice and the customers always want to pick what they know actually is awesome, right? So they out that that we actually do that and you know and one of the main aspects of Especially when you talk about these things bringing it as a service, right? We take a lot of guesswork away from our customer, right? You know one of the good example of HPC is capacity, right? So customers these are a very you know, I would say very intensive systems very complex systems, right? So customers would like to buy certain amount of capacity. They would like to grow and you know come back, right? So giving them the flexibility to actually consume more if they want giving them the buffer and coming down All of those things are very important as we actually design these things, right? And that takes some you know customers are given a choice But it actually they don't need to worry about oh, you know what happens if I actually have a spike, right? There's already buffer capacity both in so those are awesome things when we talk about things as a service When customers are doing their ROI analysis Buying capex on-prem versus versus using apex is there a point of is there a crossover point typically at which? It's probably a better deal for them to go on-prem. Yeah I mean it like specifically talking about HPC, right? I mean why you know we do have no lot of customers consume high-performance compute in public cloud, right? That's not going to go away, right? But there are certain reasons why they would look at on-prem or they would look at for example A COLA environment, right? One of the main reasons they would like to do that is purely how to do with cost, right? These are pretty expensive systems, right? There is a lot of ingress egress There is a lot of data going back and forth right public cloud You know it costs money to put data in actually pull data back, right? The second one is data residency and security requirements, right? A lot of these things are probably proprietary set of information and we talked about life sciences. There's a lot of research, right? Manufacturing a lot of these things are just in time decision-making, right? You are in a factory floor You got to be able to do that now. There is a latency requirement So I mean I think a lot of things play me, you know plays into this outside of just cost But data residency requirements ingress egress are big things and I mean you're talking about mass moments of data You want to put and pull it back in They would like to kind of keep it closed keep it local and you know make big get a get a better price point Nevertheless, I mean we were just talking to Ian Coley from AWS He was talking about how customers have the need to sort of move workloads back and forth between the cloud and on-prem And that's something they're addressing without posts You're very much in the in the on-prem world. Do you have or will you have? Facilities for customers to move workloads back and forth. Yeah, I mean I wouldn't I wouldn't necessarily say You know del's cloud strategy is multi-cloud, right? So we basically So it kind of falls in two three I mean we some customers some workloads are suited always for public cloud. It's easier to consume, right? There are you know customers also consume on-prem the customers also consume in color and we also have like del as amazing piece of software Like storage software, you know, we make some of these things available for customers to consume a software IP on the public Well, right? So, you know, so this is our multi-cloud strategy So we announced a project in Alpine and Delta cold So, you know if you look at those basically customers are saying I love your Dell IP on this on this product on this storage Can you make it available through in this public environment whether you know, it's any of the hyperscale players So we do all of that right? So I think it's it shows that you know It's not always tied to an infrastructure right customers want to consume the best time And if you need to be consumed in hyperscale or we can make it available Do you support containers? Yeah, we do support containers on HPC. We have a we have two container orchestrator We are to support we we have an opt-in a singularity. We also have a communities container options your customers both options What kind of customers are you signing up for the for the HPC offerings? Are they universities research centers or is it tend to be? It's you know the last three days this conference has been great We probably had like you know many many customers talking to us, but HPC somewhere in the range of 40 50 customers I would probably say a lot of interest from educational institutions universities research to your point a lot of interest from manufacturing Factory floor automation a lot of customers who want to do dynamic simulations on the factory floor There is also quite a bit of interest from life sciences pharmacies because you know like I said we have Two designs one on life sciences one on manufacturing both with different dynamics on the infrastructure So yeah quite a quite a few interest definitely from academics from life sciences Manufacturing we also have a lot of our financials big banks You know who wants to simulate a lot of the you know brokerage a lot of a lot of financial data because we have some You know really optimized hardware we announced in Dell for especially for financial services So there's quite a bit of interest from financial services We often think of Dell as as the organization that democratizes all things in IT eventually and And and in that context, you know, this is super computing 22 HPC is like the little sibling trailing around the trailing behind the super computing trend But we definitely have seen this move out of just purely academia into the business world Dell is clearly a leader in that space How has apex overall been doing since you rolled out that strategy what to couple it's been it's been a couple of years now Hasn't yeah, it's been less than two years. How are Howard Howard mainstream Dell customers? Embracing apex versus the traditional You know maybe 18 months to three-year upgrade cycle cap X. Yeah, I mean I look I think that is absolutely strong momentum for apex and Like we all pointed out earlier We started with you know making the infrastructure on the platforms available to customers to consume as a service, right? We have options for customers, you know to where del can fully manage everything end-to-end take a lot of the pain points away Like we talked about because you know managing a cloud scale, you know Basically environment for the customers. We also have options where customers would say, you know what? I actually have a pretty sophisticated IT organization I want Dell to manage the infrastructure but up to this level in the layer up to the guest operating system I'll take care of the rest right so we are seeing customers who are coming to us with various requirements in terms of saying I can do up to here But you take all of this pain point away from me or you do everything for me it all depends on the customer So we do have wide interest so our I would say our products and the portfolio set in apex is Expanding and we're also learning right We're all getting a lot of feedback from customers in terms of what they would like to see and some of these offers like the example We just talked about in terms of making some of the software IP available on a public cloud Where they look at Dell as a software player, right? That's also is absolutely critical So I think we are giving customers a lot of choices ever I would say the choice factor and you know We are democratizing like you said they're expanding in terms of the customer choices We're almost out of time But I do want to be sure we get to a Dell validated designs that you've mentioned a couple of times How specific are the well and what's the purpose of these designs? How specific are they? They are I mean, you know, so most of these valid I mean again, we look at these industries, right and we look at understanding exactly how would I mean We have huge embedded base of customers utilizing HPC across our ecosystem in Dell, right? So a lot of them are capex customers. We actually do have an active customer profile So these validated designs takes into account a lot of customer feedback a lot of partner feedback in terms of how they utilize this and When you build these solutions which are kind of end-to-end and integrated you need to start anchoring on something, right? And a lot of these things have different characteristics So these validated design basically proved to us that you know, it gives a very good jump-off point for customers That's the way I look at it, right? So a lot of them will come to the table with they don't come to the blank sheet of paper when they say Oh, you know what? I am this this is my characteristics of what I want I think this is a great point for me to start from right? So I think that that gives that and plus it's the power of validation really, right? We test validate integrate so they know it works, right? So all of those are hyper critical when you talk to and you mentioned the health care you mentioned Manufacturing other design We just announced the validated design for financial services As well. I think a couple of days ago in the event So yeah, we are expanding all those DVDs so that we we can we can give our customers the choice We're out of time Satish. I hear thank you so much for joining us Thank you at the center of the move to subscription to everything as a service everything is on a subscription basis You really are on the leading edge of where where your industry is going. Thanks for joining us. Thank you, Paul Thank you, Dave Paul Gellin with Dave Nicholson here from super computing 22 in Dallas Wrapping up the show this afternoon and stay with us for they'll be have more soon