 From London, England, it's theCUBE. Covering Discover 2016 London. Brought to you by Hewlett Packard Enterprise. Now, here's your host, Dave Vellante and Paul Gillis. Welcome back to London on the chilly banks of the River Thames here at Excel London. This is HP Discover 2016. This is theCUBE, the worldwide leader in live tech coverage. We've been here, Paul Gillin and I for three days extracting the signal from the noise. Vish Malshand is here. He runs product marketing and management at HPE. And Inga Erling is here from Werner, retail company. Really exciting company with numerous sites, websites and presence across Europe. Gentlemen, welcome to theCUBE. Thank you. We're talking storage, we're talking flash, all flash, acceleration, wheelhouse, we love it. What's the vibe, Vish, on the floor? You guys have been killing it in all flash. So all flash, Dave, we've been talking about this now for several CUBE sessions, right? And I think my view of the all flash and where it's heading, if that was your question here, was I think we've seen these waves of flash. Wave one back in 2012 was that performance wave, right? The likes of violin memory comes to mind at that time where there was a small number of applications that required that critical performance and then justify the cost of the flash at that time. That was wave one. Wave two is the wave we're in right now and I call it the affordability wave. In addition to the performance, we saw data reduction technologies, we saw larger capacity SSD drives, we saw improvements in power cooling density, customers reported administrative savings. So there was this economic wave that we're right now I think on the cusp of exiting to the third wave. And wave three in my mind now goes beyond the flash array. Wave three is characterized by four attributes. First is around the acceleration of applications and that includes compute, storage, network, right? It's across the board. Second is simplification. The world continues to move faster. You got to make it simple. People have other things to do than manage infrastructure. Third is managing your risk. That is access to the data, integrity of the data. And fourth is investments. Two elements of investments here Dave, how do you invest into the IT asset that you're trying to purchase? And as that investment matures over time, how does it stay fresh and stay current? Okay, good. So we're going to dig into some of those. But before we do, let's go to Enga. Tell us a little bit about Varner. Varner is a retail fashion company that is based in the headquarter in Norway. But we're having around 1,600 stores across the northern part of Europe today. We are having 12 different brands. So they're both competing with each other and also supplementing each other. And obviously online as any retailer is driving a big component of your business. Can you talk about that shift? What's happening with digital and what role Flash is playing? We're seeing that the online business is in the beginning, it will be growing quite fast. But the main reason for us to go to Flash is that the amount of data is growing. And with more and more stores, we are in retail fashion. We are having approximately more than one million unique items that we want to do big data analytics around. Customer loyalty. It's also a huge driver regarding the amount of data that you want to analyze and you want to keep track of the customers. And then all of us want to have a fast response when you're starting an application or whatever. You want to have the answer right away and then you need the right storage. How is this changing the retail experience for your customers? I think that we're still in the beginning regarding the changes. But it will be more and more that when you're going to a store, you are bringing your mobile device with you. The mobile device will be a companion. You are going to use your device to get additional information. You're already, today, you're using your phone to look at alternatives when you are in a store to compare the goods that you're looking at, what are our other stores doing? Are they a better offer somewhere else? But in the retail fashion, we are looking at what this could give you additional value or information about the product. Since we are in fashion and producing clothes, information about where is actually the garment being produced? What is the fabrics in the garment? Everything that you want to know as a customer, you want to know more and more. Why are you here talking about flash? We discover that. We ended up to have a problem regarding how far could you actually scale out with a number of disk drives? You needed to have something that had a better response time. It's more or less that you're expecting, or employees expect the same as all of us is expecting at home. When you're clicking on the up on the telephone, you want instant response. To be able to actually do a extreme, complex analytics in a database, you need to have, one thing is what could you do in the memory of this database server or the application servers? But you also need to look up something that is stored in a database, that is stored in a storage system, and then you need a storage system is extremely fast. And it's really an analytics play is why you sort of went to flash and what are you doing with the analytics specifically? Are you replacing slow data warehouse? Are you bringing transactions and analytics together? It's the amount of analytics is growing year by year and at the same time moving or building more and more on a virtualized platform that is also increasing the demands for to have a storage system that is fast. The growing amount of data in general, since we are a fashion company, we have a huge picture libraries that is adding more and more. So they're getting the de-deplication, the compression ratio that you're able to get in a flash system. So one thing is that to get the performance in packed on larger and larger databases, the dollar is also how to cover the rest of the business that we're also having in the data center. What are you seeing in terms of de-duplication and compression or data reduction rates? It's up to 1.9 in the overall when you're combining all of the different type of compression ratio that are having in the system. The other case now alone is up to 1.3, but in overall we're seeing the 1.9. And as a result, your cost of flashes at parity with spinning disk, lower than spinning disk? More or less when we did the last investment, we saw that we didn't have any really good alternatives. We're having the costs having everything from the electricity bill, the space, actually to put up a system, the number of racks that you need, and the prices was, we are seeing more and more that it was competing and we've mentioned that the new offerings that is coming with the recent gigabyte, it will completely change. It's close enough and simpler and more efficient from an energy standpoint that it doesn't have to necessarily even cross over. So in the last year we were running three different platforms and with the new offerings, we had the possibility to move on and do the consolidation again that we could have a mixed workload. You're getting more and more into the situation where you had to build one environment for the virtual environment, one for a standard clustering environment, one for databases, with all flash and the power that you're getting there, you can actually more flexible move the data around. This sort of goes back to what you were saying about simplicity, so how is Werner seeing simplicity payoffs from the flash investment? You know, in the industry that they're in, right, I mean, just coming off the Thanksgiving holiday here in the U.S., you know, we see, I think we saw a record year with online sales and I think the retail outlets now are trying to marry, right? One interesting example that we saw was a customer comes into the store. Of course, he can see and feel the material but they can take a photograph of him and they can now simulate how different looks and clothes look on him at the store with someone giving you the advice, right? So it's not yourself looking at an image on a computer screen that's there in the store, but then there's a person there that can advise you and say that looks really good on you, that doesn't look so good on you, right? Now, to do that kind of infrastructure in a store, right, really means that you've got to have very simple operations. You've got to really be able to allow people to focus on those kinds of attributes of what a customer looks like versus the infrastructure. So, but what's simpler about managing flash? Why is flash simpler to manage? It is simpler to manage because traditional systems require a lot of tweaking. They're required, the database guy, to talk to the storage guy, to talk to the networking guy. Now, many of our customers are reporting that that has gone away. So the bottlenecks? It works, it just works. There isn't the need to constantly change, tweak, respond to spikes, and flash just takes that all away. They can go focus on higher value things. Talk about the risk mitigation dimension. Now, as you're collecting more and more data, to do data analytics, for example, to drive those insights, how do you protect that data, right? And you know, we've all heard about break-ins here in the last few weeks on data security, data breaches. So that protecting the access to the data is very key. That's a managing your risk there, especially if you keep all this information, right? And then how do you protect that data over time in the wake of disasters, failures? And so, you know, as that data bank grows larger and larger, it becomes a valuable asset. You don't want to get into the wrong hands and you don't want to lose it. So, Inga, coming back to your platform, I see from the notes of you guys gave a presentation the other day, you're predominantly running, well, you're a VMware shop, running Windows file clusters, exchange and SQL as your database, right? So what kind of impacts have you seen, even at a high level as a result of Flash, roughly performance, business outcomes, anything you can share? I think that the most important thing is that we have seen is that we could simplify the platform regarding we didn't need to consider what kind of storage do we need to deploy a specific type of application? If we want to spin up a new database now, we could just spin it up. We could use the standard automatization tools that we're having and they don't need to consider are we on the right storage or not? Do we need to deploy this on a specific virtualization farm or do you need to have dedicated hardware? What is the delivery time? Today, we could just decide that we need a new system tomorrow and we can spin it up today. One of the benefits of Flash is a reduction in duplication in database copies. Have you found that that has been a significant benefit for you? Yeah, based on the overall figures, we're seeing that we're getting a huge benefit regarding the D-duplication ratio, the 1.9 ratio that I mentioned earlier. If I can add to that point, let's de-duplication, I'm thinking of just having to create copies of the database for development and product analytics. Yeah, the same copy as opposed to having to create copies. More or less we are restoring or we are actually copying data around, we are copying production data to do real-time performance testing before we're deploying new releases into production. We don't have exact figures how much it's actually saving us, that's perhaps but you're still making copies as you used to as opposed to sharing out of that single data set, yeah, and that's also due to time that you want to be sure that everything is working. Yeah, okay, but so you're IT guy, you're conservative, that's what you got to make sure. Break it first. It's an emerging use case. But it is an emerging use case and that's what I wanted to ask you, Visho. You're starting to see people, certainly in development, you've seen that. You're seeing it sort of widely or? You're seeing more and more customers now recognize that Flash enables them to do something that they could not do before, right? It's the enable new way of working. So you have a notion of copies of database, more and more customers are starting to realize that there's performance headroom in the Flash. If they use Snapshot to make the copies, they don't have to make full physical copies. So there's a data reduction element of using Snapshots. And then there's of course the protection of those different copies by fencing them using some kind of QoS software so that you ensure one doesn't overtake the other. How about in terms of what you're seeing with regard to the way in which Flash is impacting the business? So you had talked about wave one being performance and I'm actually curious as to how what an anger situation is that performance or is it accelerating apps and what the difference is there? Yeah, and I think the wave one was performance as an overriding consideration, right? And without really considering all the other elements. Wave two said, I need the performance. I need the affordability. I need the data services, right? And that made it a little more mainstream because it brought three elements together. And now wave three, we're saying, look, it goes beyond because I'm now looking at accelerating all of my applications as opposed to just my performance critical applications. In fact, one of the customers we talked to Dave said, we really did not need the performance of Flash but because of the administrative savings because of the reduction in backup windows, we moved to Flash. There were other reasons as opposed to performance moving to Flash. Now I want you to go Flash, you'd never go back. So, okay, and let's talk about the competition. Did you look at other systems? Are you a pure HP shop, HPE? We're also running other storage systems but that is for specific applications. But since we have been a long time HPE customer and the experience that we had with the earlier edition of 3PAR, we did compare the prices. And we looked at all the alternatives but the truth is part of the 3PAR was the best platform for us. When you were doing hybrid, I saw in your presentation, you were doing hybrid but it was just too cumbersome, you're moving stuff around and tearing, it wasn't working for you. Was that a 3PAR hybrid or was it an alternative? The solution we had there was running on the P9500 and the there was more or less the complexity regarding smart clearing, storage pools, licensing issues. So that's one of the benefits with the new 3PARs are all Flash that you don't have any of those statistics regarding to look into what is the lies and what is the storage pool they are running on. Just collapse all that. So let's talk about the competition here. So 3PAR came out, the interesting thing about 3PAR is you had a full stack, robust, hardened storage service stack. That gave you an advantage. Obviously EMC is a leader here, you're pure, the upstart, NetApp is now using its install base to convert to Flash. Where do you see your competitive position, Dish? So I see our competitive positioning growing because we're bringing now several elements together. If you look at acceleration, for example, right, how do we take the next wave of application acceleration? Well, one of the things we're showcasing on the floor this week, Dave, is something called 3D Cache and it's technology with HP and Intel around 3D cross points where we can now accelerate all flash arrays, right? And so that's one area of acceleration, accelerating the network. We've just announced this week as well, a whole still 32 gig fiber channel infrastructure, 25 gig ethernet on the horizon. And so when we can bring all of the elements across compute, networking and storage together to deliver that application outcome, that's a strength that very few companies can match. You're Brethren at HP, it's called Hewlett-Packard Labs. I thought it was HPE Labs. But anyway, the science says Hewlett-Packard Labs, regardless, they're working on some cool stuff, potentially to replace Flash. Are you expecting that? Are you just assuming Flash is here until somebody proves that there's something better? I think that's the rise of what some people are calling now storage class memories. On the server side, we have also seen something what we call persistent memories, right? And I think we're seeing sort of a collapse of the memory hierarchy. And we're fully expecting there's going to be a solid state 2.0 with storage class memories. I don't think Flash is dead. I think Flash continues to have a life. And the early phases of storage class memory in my mind, Dave, are going to be focused in more memory use cases, like expanding the size of the cache. Like in the case of persistent memory on the service, we're also seeing byte addressability on something like SQL Server 2016 to drive acceleration. So if we can bring now the acceleration in servers, the acceleration in storage, the storage class memory, the network, you've got to trace that end to end path of the IO across the entire stack. Okay, so, so Memrister maybe starts in the system. Maybe someday it gets cheap enough to impact Flash. My prediction is that you're going to see storage class memories, different kinds of technology start first with the memory use case. It's going to be larger capacity than traditional DRAM memory. It's going to be persistent, of course. But, you know, it'll be lower priced, right? But not low priced enough yet to go replace the Flash tier. At least not initially. Well, because of the consumer volumes especially. Over time as volumes increase, as production levels increase, you'll see it drive prices down, drive capacities up, and then it'll go down the stack that way. Would you agree we're a decade away from that, from in terms of impacting Flash, replacing Flash? I don't think it's as far as a decade Dave. You think it's sooner? I think it's five years. Okay, good. We'll follow up, you, me, and Floyer on that conversation. Okay, all right. I'd love it. Gentlemen, we're out of time. Thank you so much, Fish and Inga for coming on theCUBE. It was great to have you. All right, keep it right there. Everybody, Paul, Gil and I will be back with our next guest. We're live from the docks of London in the Atlantic, Cell London. This is theCUBE. We're right back.