Podcast Archive - StorageReview.com - Podcast #136: HPE XP8 Goes Ten Years With No Downtime

Episode Date: December 9, 2024

Brian visits HPE in northern California and has a chance to sit down with… The post Podcast #136: HPE XP8 Goes Ten Years With No Downtime appeared first on StorageReview.com.

Brian Beeler here. I'm on site just outside of Sacramento, California with HPE in one of their really great facilities. It's out, it's quite bucolic out in the country here with pine trees and everything. Really wonderful setting for this and we're here not to look at the agriculture but to get into the storage portfolio and specifically the XP series, the XP8, which is, as Rupin will tell us in a moment, don't get too excited, is one of the really great storage offerings at the top of the portfolio. One of the things that's so special about it is that it is always up. And when we get into this, we're going to talk about nines, but HPE just put out a blog a few weeks ago that's really a lot closer to 100% with 10 years of uptime.

Rupin, you guys are really excited about this 10-year commitment to full uptime. That's more than eight nines. Correct, correct. We are very proud of that. The team has worked really, really hard in producing a portfolio, a high-end portfolio of HPE, XPE product that has been up for the last 10 years. And we are claiming that as no outages for the last 10 years with none of our customers in our install base. So when we think about mission-critical storage, in the industry, these terms get thrown around all the time, sometimes inappropriately, as you're well aware. But in this case, when we talk about mission-critical storage and you saying no outages for 10 years,

I mean, those really do go together. 100%. I mean, if you look at our HP, XPE, XP customers today, most of them are running XPE on mission-critical storage environments. And customers like investment, you know, outfits, financial institutions, mission critical infrastructure, they're running XP. And they're running XP for one reason is they can't afford the storage to go down even for an hour in the whole year. I mean, it's people who require absolute no downtime are the ones who are deploying HPE, XP in a storage arrays. Yeah, it's an interesting thing because we see downtime for all sorts of

reasons. Anything like a RAM dim failure or, you know, drives or sometimes SSDs can be flaky at times. Power supplies. I mean, sometimes it's the lowest cost part that causes the problem. And obviously, you've engineered the resiliency and redundancy into the platform. Where does this fit in terms of mission criticality? I understand. What about scale? Scalability, performance, and reliability. Those are the three big mantras we have in our product portfolio, right? In terms of scalability, we scale pretty high. I mean, if you're looking at deploying a hybrid array, a combination of spinning media, SSDs, and you want to deploy a platform with, let's say, more than 2,000 disks in one

array, HPE XPE supports that. If you have a requirement for more than let's say 150 and even 190 fiber channel ports to your server infrastructure, HPE XPE supports that. You know we support up to 12 controllers. You know so the amount of redundancy and the amount of performance and scalability we have in the platform is par excellence. Yeah well so you talked about the scaling up to 12 and all the ports and all the disks, but you can go small too. This isn't just for huge deployments. You can do a smaller one too, yeah? You can. Most of our customers buy our product for the large scale deployments where they are

trying to consolidate infrastructure. But we do have a starting configuration of two controllers and two disk shelves. Our recommended solution is a minimum four controllers. That's what we recommend. If you really want to get to the reliability that we want, it's what we say is a minimum configuration is like four controllers. So if you go with four controllers then

and we're talking about resiliency, that means I assume I can take three of those out and the system will be happy, but it will be up, it'll never lose IO. It's not just up, it's up with performance, I think is your big claim here. Of course, the devil is always in the details. You have to double click and understand what the configuration is for a particular customer. Right. If you want to maintain the IO profile with multiple failures, you have to maintain a certain level of configuration. Right. Okay. So there is some devil is in the details. Sure.

Okay. But you can decide, like you said, how much uptime do I require? 100%. And how much performance do I require at those levels? And then that goes into your architecture design. 100%. I mean, I have a team of engineers reporting directly to me who actually help our customers size our deployments. And that's pretty much the first few questions we ask. What kind of performance do you're looking for? What number IOPS? You know, latency requirements. What's your replication strategy, right?

What type of redundancy do you need? And based on all those requirements, we can design a solution for you that will remain up. And we do have a history of 10 years no outages. Which is pretty strong. And I know later today, we're going to get in the lab.

We're going to try to break this thing a little bit and see what happens. We'll get into that and exploring the hardware and diving into all those components. When you think about these solutions, obviously it's part of a broader HPE portfolio. I've seen the new MP platform, which is really cool, and we've seen all the way down to the little MSA in our lab before. Those are pretty cool too. With the spectrum of offerings, how do you articulate to your customers where this sits in the portfolio and where the XP storage is really the right fit for the customer?

Correct. Our leading platform is Electra MP. Okay. We lead all our customer interactions with trying to position Electra MP. Electra MP is our own IP, and we have developed and invested a lot of R&D into that platform. R&D, I mean, when you look at that thing,

the little controller piece, I mean, that thing's brilliant. I saw that at the launch, what, a year, year and a half ago. Really cool. Little things are really cool about that platform. It's the industry's first truly disaggregated platform that you can scale now up to, I think, up to eight controllers now. But we are continuing to expand the Electra MPE's capabilities with every single release. I think we are at release four now, and we're going to continue to do more releases and expand the features and the scalability, performance, and deployment scenarios for the Electra MPE.

So we always lead with MPE. with MP, HPE XP sits at the high end of the portfolio. Because again, XP XP is a platform that customers are really buying today who want mission critical, nonstop, you know, IO profiles, customers requiring mainframe connect, customers requiring a lot of the legacy, we talked about scalability, we talked about performance. And then also we have an install base over of our customers that we want to continue to service so that's the high-end portfolio we are positioning

XP right MSA is at the lowest end of the portfolio MSA is sold a lot through the channel and you know you find MSA is a good fit for small office, remote offices, customers who don't want to spend a lot of money. We love that little thing. I mean, the amount of performance in those little systems is pretty solid. Exactly. I mean, MSA also has been a brand that has been with HPE for more than 20 plus years. Wow. You know, and although the the product has changed and has evolved over the years, but the brand has remained there starting with the ProLiant.

It's it actually was a, you know, you know, it was very well positioned with the ProLiant. Right. And even today, our attach rate of the MSA is very high with our ProLiant servers. So we continue to have that in our portfolio. Like I said, it's a screamer in terms of performance. But it's just a tiny little box that delivers the price and the performance ratio that our customers want. But again, Electra MP is an amazing platform. It's industry's first disaggregated platform. It's based on the three-part DNA, as we were discussing earlier. It's industry's first disaggregated platform. It's based on the three-part DNA,

as we were discussing earlier. It's really our go-forward platform in terms of the mid-range and scaling up and down both. But somehow, we always have had these customers who really love MSA at the lowest end of the market and XP at the high end of the market. And we want to continue to make those customers remain happy customers. So you talked about XP for mainframe, and that makes a lot of sense to the largest organizations in the world that have that as a fundamental part of their business. But we shouldn't think about it just as mainframe, right? There's this whole open system side of the world. There's a couple other neat features built in too that are worth mentioning. But talk about the open systems play because I think that's as critical for your customers.

100%. We have a huge install base in open systems, 20% are mainframe. So HPE XP continues to do very well in open systems as well as nonstop mainframe, etc. So with XP, to get to this 10 years of no downtime, it takes more than just deploying the physical hardware. HPE is able to leverage additional testing, validation, services, software, insights that really helps you differentiate and take that product to this 100% level that is quite unique. Talk about that a little bit. 100%. 100%.

HPE XPE is a very unique platform. We have almost 100% attachment with our services. So pretty much every storage array we are selling today and we are helping our customers install it, deploy it correctly. And then we have a nice service plan associated with that particular installation, right. And in fact, last two, three years, we've had a huge uptick in the GreenLake, you know, GreenLake umbrella, along with that services umbrella. So we have a lot of customers who are buying HPE, XPE as a classic GreenLake solution, where we are coming in and taking care of all the infrastructure for that customer,

whether it's from end to end, we help the customer, we work with the customer and take the customer through the whole life cycle of compute, storage, networking. And that part of the business is growing like weeds. And we love it because we love the GreenLake business because we can take care of our customers really, really well when we have the GreenLake umbrella. We can fully control their infrastructure, arrays coming in, data migration, arrays being phased out, maintenance of the arrays, maintenance of the drives, maintenance of the firmware. So that is something that actually, that's why I love Antonio's GreenLake strategy.

It's a winning strategy for us. Whether you buy as a service or whether you want to buy as a CapEx model, the GreenLake strategy. It's a winning strategy for us whether you buy as a service or whether you want to buy as a capex model. The GreenLake umbrella really is something that's very much resonating with our customers. Well, I mean HPE was early money on that methodology on going to market in the management of systems like that. I mean this has been what four or five years now I think since you started that. So we were the number one in coming out with a hybrid, you know, strategy and the Green Lake strategy, right? I mean, you're seeing that in the cars business, right? I mean, last two years,

everybody was going towards EVs. Like it was fully plug-in, fully plug-in. People wanted a fully plug-in all-electric car. Well, happened this year electric cars are sitting on the lots and customers want a hybrid car the customers realize you know what I have that anxiety of running out of you know battery you know in the middle of a trip yeah so you what selling right now is you know hybrid cars so that same thing is you know is happening now in the IT landscape. HPE's hybrid strategy is right now the winning strategy because customers went to the cloud and they realized,

oh my God, I'm spending so much money in the cloud. I went to the cloud to save some money. Yeah, it was easy, but it was not inexpensive. Correct. But what they really liked is the cloud operational model. So, with our new platforms, Electra MP, we have the cloud operational model with GreenLake, DSCC. But you don't have to spend like you were ending up spending in the cloud. So I think the hybrid model is the way to go.

When I talk to customers, they have all kinds of applications running their business. It's not just one application. They want to run their most mission-critical applications on something like HPE XP. Well, they also, if I know storage admins, they want to see the stuff that they're running their mission-critical apps on, which is why the on-prem hardware with the management wrapper on top is a pretty neat combination. Correct. And then Electra MP is an amazing platform.

It's contemporary. It's disaggregated. It's the new way of thinking about storage. And I think with the cloud operational model in GreenLake, it's going to be, it already is a mega success and I think it's going to continue to ramp up and really become our flagship and our spotlight array going forward. But for our customers requiring more of a classic array, traditional hybrid array, something that's highly performed, high performance, super scalable, very, very reliable, mission critical. HP, XP fits into that part of the portfolio. The one thing I think is kind of neat about the MP, and I don't want to spend a ton of time

here, but you brought it up, so I will. The block is great, but the same platform can be used for file, which gets back to sort of what you're talking about, a disaggregated sort of universal platform for software going forward from HP, which is pretty neat. 100%. You know, it's the concept of personas, right? You can buy a single hardware platform and deploy different personas, whether it's block, file, or object. Again, Electro MP gives you

that capability, right? Yeah. Which I don't think any other competitors allows you to do. No, it's pretty neat. So talk to me about migration because I know that's another thing that's a big deal. When you look at this, especially the mainframe business, the customer list is somewhat well defined, right? You understand who's out there, who's using these things.

And the lift to go from one system to another system is tremendous. But you guys have something that I didn't quite appreciate until we talked through it a little bit more about the ability to front-end additional storage, and not just mainframe, but on the open system side too, right? And that one's pretty neat to let customers take, not take their time, you know, to be slow about it, but to be programmatic about how a migration might take place.

Tell me about that a little bit. Yeah, and, you know, the external storage feature on the XP platform is probably, you know, the only array in the industry who has such an extensive support for external storage. Another interesting part of that virtualization capability that the XP platform has is that code has been up there for, I think, 20 plus years. So it's very reliable. The feature set and the support matrix of external storage is deep and wide. And it has a long runway.

So in terms of whether you want to migrate, the simplest way to migrate is configure that third-party array as an external storage to XPE, and it absorbs and is able to present all that data to you almost instantaneously. So that particular feature, you're absolutely right. I think we should talk more about it because I think it's a very powerful feature in the platform. Well, as I think about the investment and making a change,

I mean, there's a lot of fear about that or uncertainty. And if you can mitigate that by saying, we can set up in front of whatever you have today, take those shares, the LUNs, make those available through a sort of pass-through. But in the meantime, you can also begin the migration process of offloading that data while still serving it. That's pretty neat. 100%. And I mean, customers, I mean, data migration

is, I mean, think about yourself. I mean, you know, your family photos and storage. Even managing that has become a challenge, right? You have to think about a data protection strategy for all your kids' photographs and videos, right? You know, house burns down. You don't want to lose those photos. Those photos are probably the most valuable asset you have in your home, right? So what are you thinking about, right? Like, you know, Apple,. So what are you thinking about?

Apple, if you cross that one... It's another $9 a month or whatever. After you cross one terabyte, you have to go into the business class, which is way more expensive. And also the fact that it takes so much time to upload and keep all those copies of data. So data migration has become a problem not just for the enterprise. It's become a problem for everyone, right? Sure.

And to your point, the HPE XP external storage feature makes data migration literally painless, right? You don't have to bring a third-party box. You don't have to bring a third-party company. You don't have to go and do any complex fiber channel zoning. The platform can actually take care of both migration of your data as well as supporting your older arrays while the migration is taking place. Well, and it's important too because your team, right? We all have limited IT resources.

You've heard it, I'm sure, a million times from your customers. Anything you can do to make their lives easier, especially at the mission-critical level. This is not a little baby NAS for a little file share. This is the real deal. With that, I think there comes... Many think there's an implied complexity to the systems. And not to say it's not complex, because surely it is, but the management doesn't have to be complex of these things. To be honest, I think the management, even if it is complex, and I'm not saying that HPE XP has a complex management platform. It's actually pretty intuitive. I won't say it's the most simplest.

But for the storage admin who has a large infrastructure to take care of, he actually doesn't mind spending some time with configuring the storage. Because they want it the way they want it, too. Exactly. They want it the way they want it, and they want to get it right. Yeah. There's a lot going on there. And yes, the storage admins are very specific about how they want these things deployed. But once it's done, I mean, you're up and it's running for decades or more, as you've said. Like I said to you earlier, our platform has pretty much 100% services attached.

So we have a very talented services team along with a product R&D team that makes sure that this product is installed properly and it's serviced properly. So on the spec sheet, you list, what, eight nines with XP8? Correct. But in reality, you're delivering above that, which is nice. The level of nines, the six nines, the eight nines, whatever it is, what does that mean to your customers? I mean, we can all math it out and it would be 12 seconds over some period of time. But financially at that level, these are really serious issues where even a little blip is expensive.

And as HPE, if you're not delivering on that quality of service, you've got a financial risk to that commitment of uptime as well. Can you talk about the real impact of what the nines means to a business? Think about train systems being shut down because of some software update or storage not available, you know. Think about, you know, a trading platform, you know, which goes down for even a few hours, you know, middle of the week. I mean, the impact on global markets would be unbelievable. 100%. You see what happens now, you know, you have some blip and some blip and you have a stock market meltdown. So I think that's what I was talking to you earlier, that buying an HPE XP storage array is almost like buying an insurance policy.

When you want to make sure that you can't afford to have your storage go down, that's when you buy. It's not like something is going to happen, but you buy that platform so that you know that it's not going to go down. So those are the types of customers we have on the platform. Customers who just can't have their business go down. It's not just about losing some X amount of dollars if the business goes down.

Some of our customers will actually go out of business if there's a big outage. They will literally just go out of business. Which seems consequential, I would say, right? And so that's why you're here and why that exists. What do you see looking forward with data? Because, I mean, I think we think about, like the mainframe business, I said, we sort of know who uses those. But is there an increase in demand for this class of storage, this Michigan critical storage, or is it kind of finite?

Like, what are we looking at going forward? When we look at all the IDC data, and that's what we have a picture on, right? I mean, I think the high end of the market, it's relatively flat to a maybe low single-digit decline. Okay. So contrary to a lot of people, you know, prophesizing that, you know, the high-end market is just going to go away. Why would it go? I mean, it doesn't make any sense. It doesn't to go away. Why would it go? I mean it doesn't make any sense. It doesn't to me too. So this particular part of the market has been pretty stable and it's like flat to maybe

you know low single digit decline. So it's net net it's not going away anywhere. It's a multi-billion dollar market. I think last I saw it was like a four billion dollar market. So it's going to sort of remain that size, the market, going forward for the next, I would say, at least a decade. I think you're going to really appreciate the platform. As we go to the lab, you'll see the capabilities, how amazingly powerful the platform is. And I hope you enjoy the demo.

Let's go.