>> Announcer: Live from Vancouver, Canada, it's theCUBE! Covering OpenStack Summit North America 2018. Brought to you by Red Hat, the OpenStack Foundation, and its ecosystem partners. >> Welcome back to theCUBE, helping to extract the signal from the noise. Here at OpenStack Summit 2018, I'm Stu Miniman, my co-host for the week is John Troyer. Happy to welcome to the program atmail, which is an email as a service company. We have Jay Sil who's the European Sales Director, and we have Jason Brown, we'll call him JB for the rest of the interview, is a Solutions Architect. Gentlemen, thanks so much for joining us. >> Thanks very much for having us here. >> Alright so Jay, email is a service, tell us a little bit about the company and the state of email, haven't Office 365 just taken everybody over? >> Jay: Well, so most people don't want to talk about email, but it's still essential. So atmail is a 20 year old company, we are probably one of the largest pure plaid, white label email providers in the world. We have about 170 million mailboxes out there in the wild. But we provide not to end-user businesses, we service the service provider and telco market. So a lot of our customers you would've heard of, we're more the brand behind the brand. So, we provide those email to their end-user subscribers, but it is very much the telco ISB, that is upfront that you would hear about. >> Yeah, excellent. There's been a discussion we've been having at this show a lot is, OpenStack itself is kind of something that gets in there, the telco and service providers, big place so, JB, tell us a little bit about your role and bring us into the involvement with OpenStack. >> JB: Sure, so I'm the Solutions Architect for atmail, I kind of help bridge the gap between the technical and the non technical, I help Jay out with explaining the technical details to the sales team and then bring back the non technical details of feedback that Jay gets and we get from our customers, into development and operations, so they can actually improve the product in a way that's fitting. And so, we started with OpenStack a few years ago, through a partnership with DreamHost, here in North America, to move from, we kind of had a traditional email, like a hosted email solution, or an on-premise email solution, but it wasn't a true Cloud solution, and so, took a big step back, looked at our architecture, what it actually looked like, what it needed, and it just turned out that OpenStack was the best direction for us to go to make that move. >> JB, can you clarify, when you say a true Cloud solution, what did you mean by that? What were kind of the requirements and what did that? >> So we had, for years, we would just take our on-premise solution, and we would run it in a data center that we had a rack in, we had 40 U's worth of servers, I was the guy at the time that was responsible if something went wrong. I got a call at three o'clock in the morning to drive to Spokane to go to our data center to fix something, replace a hard drive, or do something like that, and that just was, it didn't scale horziontically or vertically to be honest. That was just limited to what we could do with it, and so we really wanted something where we could save the cost by distributing a load as we needed it, and I think that's really the difference, is you can spin up instances for front end or spin up an instance for a back end, whatever you actually, whatever resource you actually need, you can spin that up as a service, in a Cloud infrastructure, whereas you can't really do that as easily or as cost effectively on bare metal. >> Jay so, I want to bring it back to the business. Your customers, what does OpenStack mean from them and the ultimate end-user, I don't think I've seen emails that say, "Sent to me via an email service powered by OpenStack". But, walk us through what that means for the business and your constituents. >> So there are both commercial and technical benefits. If I look at the commercial benefits first and foremost, what OpenStack allows us to do is to provide a solution, quickly and efficiently. The first thing that people want from email is they want a stable, robust service. It's a bit like turning a tap on at home, and it getting clean drinking water. You really don't give it a second thought, its only when that tap stops working and its not coming out properly, then you think about it. So first and foremost, our customers want a stable, mature, reliable service. They also want to make sure that it's secure. And that allows us, the OpenStack initiative that we've undertaken allows us to achieve that. The commercial other benefits that we obtain from that is being able to reuse our cost base, or controlling our cost base. As a result, that's passed on to our customers. So they can then, not only mitigate their risk, but they can control their costs as well. From a technical point of view, I mean, JB can touch upon some of the technical benefits, but one of the things that we found, because we are a small vendor in terms of the DevOps team that we have, what OpenStack allowed us to do was to gain from the knowledge that the community had, and really benefit and accelerate our solutions market. And when you talk to some of our DevOps guys, the first, and, well, foremost thing that they say is that we couldn't have achieved this without the help and support of the engineers and the OpenStack community. So the depth of knowledge out there really helped us accelerate those services. >> That's great, is the fact that it's OpenStack, seems like at this point, one of the themes that we've been talking about is OpenStack, ubiquitous, mature, a lot of talk here about containers and other things, but the Stack itself is well known and mature, that seems that that would also have a impact on, something that telco understands, right? It's a well known Stack, yeah. So JB, this is your first time, you said that this is your first time at a Summit. Kind of curious, before we dig into kind of maybe what your Stack looks like, OpenStack looks like, what did you think of the Summit, the level of kind of conversation here, the sessions and that sort of thing. >> So far it's been fantastic. I've had a complete, not a 180, but there's so much here that I'll be able to take back to our DevOps guys and our QA guys, we're looking at the zool stuff really heavily, the CICD stuff, just a huge benefit that'll streamline all of our development and testing and then pushing that to market will be huge. >> Anything specific, 'cause one of the things we've look, there's a number of CICD offerings in the market today, what specifically about zool, because you're using OpenStack that it makes sense to fit somewhere. >> Yeah I liked it, it fits with OpenStack really well, I like its level of maturity, and I like the gated looking at the future as opposed to looking at the past, or looking at the present, for your testing, specifically. >> Gotcha, that's interesting, yeah. Can you talk a little bit maybe about your so your Stack is a, so it sounds like, well yeah talk a little bit about the OpenStack, your OpenStack deployment in terms of there's a lot of components, are you using kind of the core components then? And anything else that interacts with the other theme here, right, is OpenStack has to talk to a lot of other systems. >> So we use a pretty, we use the OpenStack storage module and the networking module, and I don't know all of the little names to all of the little pieces, but we do use the storage and the networking. The networking was a really big help for us because we were actually able to offload some of the system load into the network layer moving into OpenStack, whereas before we would have, with an email system you have all of your actual email traffic, or your high map traffic, can create a significant load, by being able to move some of that load into the networking layer, we're able to provide a better customer experience because all of those edge services aren't as taxed, and so when the user goes to check their email, or send an email, they're not waiting because of a high level risk, and if you see this, especially if, when, something goes wrong in a system, 'cause they're systems, and things do happen, and so when that happens, the time to recover, is faster on our back end and the overall the way that's presented to our end-users is much better for us. >> John: Much better business benefits, yeah. >> Jay, have to think in the regions that you play, kind of the governance and compliance, something you need to worry about, also it's May 2018, so I have to ask you about GDPR, and how that fits into your business these days, so. >> Jay: Absolutely Stu. So, GDPR comes into effect this Friday, we've had a team dedicated on working on that, make sure that we are compliant, obviously our telco users, service providers, rely on us implicitly, to make sure that we are fully compliant, and I can assure you that we are. We have seen a number of high profile breaches of other offenders, it's not something that we want to have an experience of, so we have worked diligently, in order to make sure that we are fully compliant. >> Any commentary you want to share on the security these days too? As people always, governments asking for things, hackers, it's a complicated issue. >> It is, and it's interesting because email, I think, represents the largest surface area of attack, in any organization. You can get from a CEO, to anyone in the organization, via email. That's how powerful it is. And again, as we were talking off record earlier, it's not something you give an awful lot of thought to. Email is like turning on a tap at home and clean drinking water comes out, you don't give it a second thought. But when it stops working or there's an issue, than when it becomes a problem, and you could regress back into the dark ages, because you can't do business, you can't send that message, you can't communicate or connect to the audience that you want to. So, yes we have a lot of issueS around that that we need to make sure that we are fully on top of, our aim is to provide a stable, mature, reliable and secure service to our customers and their end-users. Security is something that we take seriously, as do a lot of other vendors, but it's something that is always constantly changing and evolving. By the time the latest attack comes out, and you've checked that you are covered, the next one has come out. And we've seen a lot of attacks over the last few months that come in waves. We had one acry last year, that really hit UK and Europe hard, as with other regions, and I'm sure there'll be more coming out soon. >> JB, containers, well, secure containers, one of the topics of conversation here, containers in general been a big topic, Kubernetes, how are you all looking at that application and orchestration layer? >> Containers with an email system are kind of tough. Security is a big reason for that, and its not that we can't use containers, but by the time you take a container and wrap all of the security around it, and everything that you need for something you would use with an email system, it almost negates the benefit of using the container to start with. >> John: Gotcha. >> So we're constantly looking at other ways that we can take advantage of that, and Koda I think today, just released their version one of their solution, which secures it down to the into the actual core of the system, and so that changes the game a little bit, on what might be possible now, not having to worry about some of the security issues that we are concerned with. >> Right, so, but even now, your Cloud portability strategy per se, is your app runs it's on an OpenStack context, with OpenStack configuration, you run I think at least two on two different instances of OpenStack, so that's part of your, you are multi-Cloud in that sense. >> We are, yes. >> That's great. >> And that actually made it really, the move into our EU data center was so much smoother, because of our experience with OpenStack on our initial deployment. We were just able to just launch it and go. >> Stu: Alright well, what I want to give you both just the final word is to, your takeaways here at the show so far, being first time attendees. >> So, from a commercial point of view, I mean the networking has been tremendous. I've had conversations with people over email or over phone, that I've actually met face to face here and made that connection, so for me as a sales person, those networking events et cetera, have been invaluable. What I also like about the show itself, and the community as a whole, is that there is this openness and there's this willing to share ideas which you don't always find in other arenas, it's much more of a closed, well I'm not going to tell you what I'm doing because its a trade secret or its going to give me advantage, whereas here it is very open, it is, we want to collaborate, we want to share, and that's been very refreshing from my point of view. >> The community is a big part of it for me. All of my work in developmental operations has been from the OpenSource community so, to come back and see that thriving and pushing this forward the way that it is, its just so reassuring. >> Well Jay and JB, we really appreciate you being open with sharing your story with a practitioner so, thank you and congratulations atmail for all that you've done here in the community. For John Troyer, I'm Stu Miniman, much more coverage here at the OpenStack Summit 2018 in Vancouver, thanks for watching theCUBE. (electronic music)

>> Announcer: Live from Las Vegas, it's the Cube. Covering VMworld 2017. Brought to you by VMware and its ecosystem partners. >> Welcome back to the Cube. Our continuing coverage of Vmworld 2017 continues. I'm Lisa Martin with my co-host Stu Miniman. We're excited to be joined next by Jason Brown a Cube alumni consultant and product marketing for Dell EMC ScaleIO. Welcome back to the Cube Jason. >> Thank you for having me. >> Good to have you here, so day two of the event, lot's of announcements, lots of buzz. Talk to us about ScaleIO. What's the current state of the business. >> Well, it's actually really exciting right now. We're doing really well. We're seeing great customer adoption. We're seeing massive petabytes of ScaleIO deployed in data centers, and were here at the show really to talk to you about customers for ScaleIO for Vmware. 'Cause everyone here's above E10, obviously, they're doing awesome. We love it. They're doing great. But there's some differences and similarities between the two products that people get confused about, so we're here at the show really trying to help, you know, ease confusion, talk about how it's like peanut butter and jelly, right? Some people like peanut butter, some people like jelly but most people like 'em both, so we're just trying to help people out and understand when to choose which and sometimes it's both. >> Alright, Jason, I've got a history watching ScaleIO since before the acquisition, you know, service providers that usually kind of fit their model a little bit more than VSAN, so when I think scale, I tend to think ScaleIO. I interviewed ADP yesterday. Big customer, rolling out like 30,000 nodes of compute with VSAN. So, scales >> Yeah >> not only one piece of it. Maybe, help us kind of understand some of the, you know, of course there's going to be places that overlap, but what is the, you know, kind of ideal ScaleIO customer, what are they looking for, and how's that differ from the VSAN? >> Sure, so in particular if you're looking at ScaleIO for VMware, there's a few things you need to understand. First and foremost, with ScaleIO we're talking about consolidating resources across the data center. So we're talking data center grade software to find storage which can run in a hyperconverged model or not. And that's really key differentiating, 'cause if you look at these enterprises, especially, these large enterprises that built an IT organization of past 20 years, right? And so when you introduce HCI to them, you're transforming the architecture of the data center but also the IT operating environment. And that's scary for a lot of people who have spent millions of dollars having a server team, a network team, and the storage team. So one of the key things for ScaleIO in a VMware environment is, if you want to transform the architecture to software defined, but preserve that IT operating model, this two layer deployment, we call it, you can do that with ScaleIO. But on the flip side you can also do a more modern architecture with hyperconverged as well. So you can get the best of both worlds. So whether today you're ready to go all the way with the service providers, they'll go hyperconverged, you know out of the gate, but enterprises usually start more traditional and then move to that hyperconverged and ScaleIO provides that pathway to get there. >> Yeah, bring us inside those customers a little. 'Cause I've talked to a couple of very large customers of ScaleIO actually, did a case study at Citi and Citi told me, internally, we're just not ready to go fully hyperconverged. >> Jason: Exactly. >> So they kept that. They're massive scale. Talked to a large global hospitality company that, once again, looked more at kind of the storage usage of what they're doing so, I mean hyperconverged VSAN seems to be having, you know they've got 10,000 customers, they're all in that-- model. >> Exactly. >> So, what is it that gets a customer ready for that? What kind of pushes or pulls them towards being ready for, you know, embracing? >> Well, I think it's understanding your business goals and your desired outcomes. So with something like ScaleIO you're looking at simplicity in the data centers. So you're looking for scale, you know, not tens of nodes where traditional, I hear this said that traditional VSAN deployment is eight to 16 nodes, 'cause they're you know, VMware's everywhere, right? There's a lot of ROBO, SMB, VDI, use scales right there, and that's not really where ScaleIO plays. ScaleIO is about data center, so Tier 1 application, databases, data analytics. It's looking at things like containers and microservices, Splunk, NoSQL. Applications like that. So when you look at those types of applications and workloads, you have to understand that your scale will probably go from tens to hundreds of nodes. Your performance may go from a million IOPS to tens of millions of IOPS. You may need six nines availability 'cause again, you're running in the data center. Customers are replacing their SAN arrays with ScaleIO. So you need all that enterprise class, data center grade functionality with the scale performance and flexibility, the key thing is flexibility as well, if you want to run multiple workloads on a cluster, you need to be able to support VMware, Hyper-V, KVM, Linux, Windows, so and ScaleIO enables all of those things. And therefore, that's why when you look at your business goals, your business ops and what your data center looks like, you need to understand that functionality. Then you decide okay, is it going to be VSAN or ScaleIO or is it going to be both, 'cause I have both of those use cases there. >> So you talked about VSAN and ScaleIO, peanut butter and jelly. Michael Dell on main stage with Pat Gelsinger said VMware and Dell EMC are like peanut butter and chocolate. Both, all good flavors, in my opinion. I'd love to hear an example though, of where, like to your point, before I asked the question. We just had the CTO of Dell EMC storage, speaking with Stu and I a few minutes ago and one year post-combination, and he said customers are starting to understand now the value of Dell EMC-- >> Yes. >> Together. So with that, you know, a year later and customers now understanding the value proposition of this company that now also owns VMware, how much easier is the conversation, you know, away from VSAN verses ScaleIO? I'd love to understand where are you seeing where they both, those peanut butter and jelly sandwiches play together. What are some of the maybe industries or key use cases where a customer would need ScaleIO and VSAN? >> Sure. So if you think about financial services, Citi as Stu mentioned, one of the larger ones there, definitely plays there, in healthcare there's a few large big partner network companies that have come together to be successful there. Telco, Verizon, Comcast, right? Not only just private Cloud but public Clous as well, so when you look at your data center, you got to look at the whole thing. So, for your VDI, your ROBO, your SMB and maybe for a few of your enterprise applications that only need you know, 50,000 in an IOPS performance for your VMs then VSAN is going to be great there, but then you look to the other side of your data center and you've got something like SAP, you know HANA, I think any other, in fact, ORACLE, etc or you're looking to build a private cloud of hundreds of nodes, well that's where ScaleIO is going to sit. Over in that corner, you know? So, it really is understanding what your workloads are and where they play. You know, it's important to know too that for ScaleIO our primary use cases are array consolidation, so you've got silos of arrays in your data center and you want to stop managing silos of arrays, and you want to bring everything together into a single resource, a single cluster, boom, ScaleIO. You want to build the cloud environment whether you're a service provider building a public cloud like Swisscom for example, who built a public cloud based off of ScaleIO, or a private cloud like CitiGroup for example. It's pretty much a private cloud; mix of array consolidation as well. And then something like a gaming company that we've worked with where they are doing this next generation DevLogs containers, microservices, well ScaleIO's great for that too, 'cause it has the flexibility to start small and grow and support the various things that they need to be able to deploy their applications 32% faster. So you know, it really encompasses the whole data center. >> Yeah, a bunch of interesting points that I want to unpack a little bit there. Specifically, you're talking about all the new applications and the new technologies that people are doing. One of the challenges most people have, you know, the stack we've been using, I think, for my entire IT career is, you know, we spend what, somewhere between 70 and 90% of our time keeping the lights on. >> Jason: Yes. >> And the wave of kind of software-defined, you know, all of these type things, supposed to be, we need to simplify our environment, you know and, therefore I can take those resources and reallocate them, retrain them, put them on cool new things. What are you seeing from the customers, you know, just organizationally from what happens to the storage people as well as how do they take advantage of some of these tougher things like application modernization? >> Good question. Good question. So, you know it depends on the company right? There are, like you said, there are some customers that want to keep them separated and that's perfectly fine you know, there are tools that you can use with ScaleIO so that you can manage the storage independently of the compute. But then you've got things like our tight integration with vSphere, where the VMadmin can manage the storage as well. So, it depends on the preferences as well as the maturity of the organization and the skillset of the folks that are managing it as well. If you can have a storage admin become more agile and be able to manage the compute and the VMs as well then perfect. They become more generalists, right? We've talked about how these specialties becomes more generalists in these types of HCI and NextGen environments. So if they have that skillset then perfect and both ScaleIO and VSAN can enable that. And then if you're looking at app modernization, you know what do you need from an infrastructure storage perspective to achieve that, and how can you enable your application developers access that storage even faster? And that's really was ScaleIO does with the whole automation points behind everything. With, be able to add resources on the fly, remove resources on the fly, reallocate on the fly. So being able to be flexible for what they need when they all of a sudden are ramping up a new application is really critical. >> Yeah. I guess, I'm wondering if you have any specific examples. One of the critiques if you talk about, you know, storage, admins, fast is not something that usually, you think of. Flash is fast and everything like that but, how do we keep up with the pace change, how do I move things? How does ScaleIO help change that equation? Even just specifically for storage? >> Well I think that in order to be able to keep up with that change, right, it's about, as you said, simplifying their job and making it easier. So, if you've got the tools and the, just the functionality in the product itself, to be able to help them learn faster, be able to press a button as opposed to being able to allocate an array group and (murmurs) things that have an architecture, that makes that be able to achieve that as well, that's really how you do it. You know I haven't talked to any storage admins lately, unfortunately. So I can't give you a specific example, but that's really what we see at kind of the one on one level. >> And from a buyer's train of perspective, so much has changed and shifted towards this C-Suite. When we look at things like data protection, we, you know, some announcements about that yesterday, storage, and you said you haven't spoken with storage admins in a while. There's a lot of data that show that data protection storage isn't an IT problem, it's a business problem. So how has the conversation now with Dell EMC with respect to whether it's ScaleIO or whatnot, shifted upstream if you will, talking to more senior executives, rather than the storage guys and gals that are managing specific pieces? Tell us about that-- >> Sure. >> Conversation and maybe cultural shift. >> Well when you talk to any C level executive, what's the top of mind, right? Security, saving, cost savings, budget, right? So when we're talking to executives, where they talk about data center transformation, how software defines storage and enables that both at the architectural level and at the IT level, but also about how we can make their business easier to run and how it can save them money. so if you're able to get all this great flexibility and scalability and all this you know, performance, but then be able to preserve the features that you need, like compression and snapshots and being able to connect to your data protections suites as well? So if you can tell them all that and say hey and you know what, we have customers saving 50% five year TCO by doing that, without needing to do data migration or tech refreshers anymore. They're like alright, sign me up. Because you have to understand too, when you talk to them, they don't need to go buy an array the next day, and spend a couple million dollars they maybe be will be able to utilize in the future or not. They can start very small. Three nodes, four nodes, and have this pay as you go licensing so they love that as well because it grows on their terms. Not on our terms, on their terms. And that's really important for you know people that in those C level suites that are trying to maximize the efficiency of the business. >> Alright, Jason, one thing's when customers buy into a solution like this, it's more of a platform discussion these days and of course one of the things they're looking for is where are you taking me down the road? So it's great, here's what I can do today, one of the things I love this whole wave of it, is, you know, upgrades and migrations were like, you know, the four letter words for anybody in storage. >> Dirty words. >> And I said, you know, when we have a pool of resources and I can kind of add and remove nodes it was like, oh my God, that was, we conservatively estimated like five years ago that 30% of the overall TCO was based on that alone and. Wow. Scrap that. Last time you're ever going to need to, you know, migrate once you get on this platform. But, I want you to talk to us a little bit about, you know a little bit, kind of the vision and roadmap. What are >> Sure. >> You talking to customers about. >> Absolutely. So, you know with a product like this, it's constantly evolving and innovating so when we talk to customers about what's in the future, well you have to first be thinking about data services. Data services are always very important and with ScaleIO, you know, admittedly, we're a little short on some data services because we more focus on scalability and performance and making sure that we have a six nines architecture. So, the first and biggest thing that's coming very soon, if you were at Dell EMC with ScaleIO is compression. So being able to, you know for your block storage workloads, being able to maximize the efficiency of your storage even more with some in line compression? Very important. So we're doing that. We're also enhancing our snapshot's functionality so that, you know when you talk snapshots and SDS, you know, you compare it to an enterprise array, probably not up to snuff. Well what we're doing now with our snapshot keeping in relation to ScaleIO is we're actually going to have them be better or even much better than something you'd find in like an all flash array. You know, where you can have you know, thousands of snapshots in a v-tree and things like that. But it also goes to hardware as well. 'Cause there's always hardware, right? And with the innovation within Dell EMC with Dell PowerEdge servers with our friends in CPSD, we're able to innovate a lot faster with ScaleIO and SDS. So, 14G was announced. Well ScaleIO's going to be one of the first products within Dell EMC through our ScaleIO Ready Node to support mV dims and MVME. So as you know we support MVME today, one of the few software device storage platforms out there today that supports it, in a roll your own server model. With the Ready Node 14G coming out later this year, with the ScaleIO Ready Node, immediately out of the gate mVdim and MVME technology in a ScaleIO Dell EMC hardware product, 'cause it's already you know its Dell PowerEdge servers and ScaleIO software. And then helping our management keep our management keep (murmurs) as well so, introducing VVols for our VMware customers, being able to provide something called AMS which is our automated management services for the Ready Node so that you can deploy, configure, manage, upgrade, not only the storage software but the firmware as well as the EXS hypervisor all in a single button, in all a single interface, so we're doing that as well. So it's all about, you know, taking advantage of NextGeneration functionality from the hardware perspective, simplifying the management, then introducing critical features and functionality that our customers have been asking for. >> Just to make sure I'm 100% on this, things like the data services, that's software, so everybody that's got it today, will be able to upgrade it. Obviously the next generation of hardware always helps along the way, but you know, you manage those a little bit separate even though you want to handle both of those vectors. >> Yes, exactly. So when you upgrade to ScaleIO.next when it comes out you'll get that feature functionality. Now there's a few things you need to understand, right? You should have Mvdims and some type of flash media to support it. >> Stu: Sure. >> Because you're trying to maximize scalability and performance while providing these features, there's some dependencies there. But yeah, out of the gate, those features will be available. That's why it's called software-defined storage. It's all in the software, all this world of goodness is. >> Okay so take me upstream. Lot of new features, functionality coming out; what are the new business benefits if I'm the CEO of Swisscom, that I'm going to be able to achieve from that? >> Well I think definitely increased performance. Definitely increased efficiency of your storage with things like compression and snapshots. Now, if you're able to compress that data, get more out of your system-- >> But what kind of like, in terms of TCL. How am I going to be able to reduce. >> Oh, well. >> What are the factors of-- (grunts loudly) >> You know, we haven't run the numbers yet, but you know, the fact that we already can achieve 50% TCO, it can only get better from there when we're introducing these types of features, where you're maximizing efficiency, so, we expect it to bump up a bit. We're hoping we can work with you guys to get some good numbers that come out of it. >> Excellent. So continued strengthening of those-- business outcomes is, >> Yeah, that's it. You know, making sure, >> what you're talking about. >> Makings sure that the customers that want to move to software-defined storage in their data center, are able to achieve that in the most seamless way, and be able to reap the benefits. >> Fantastic. Well Jason, thanks so much for sharing your insights what's happening, um, peanut butter and jelly. Makes me hungry. I think it's time for lunch. >> It is lunch time, yeah. >> We thank you so much for coming back-- on the Cube. >> Thanks for having me. I really appreciate it. >> And for my co-host Stu Miniman, I'm Lisa Martin you are watching the Cube live, day two of our continuing coverage from VMworld 2017. Stick around. We'll be right back after a short break. (electronic music)

>> Announcer: Live from Las Vegas it's The Cube covering Dell EMC World 2017. Brought to you by Dell EMC. >> We're back. This is Dell EMC World. We are The Cube. We're here in Las Vegas, Dell EMC World. Three days of back to back coverage. My name is Paul Gillin, and joining me right now Jason Brown, from product marketing at Dell EMC and Narasimha Krishnakumar, director of product management at VMware. And we're going to talk about software defined storage. What is software defined storage? Well everything is software defined these days, right? Our networks, our servers, our desktops, and now our storage arrays as well. We've been talking a lot about hardware storage at this conference today, but maybe we won't need the hardware so much the same way as we used to. I don't know. What is it? We'll start with that. Narasimha, what is software defined storage? >> So software defined storage is managing storage through software. Where you have a data plane that resides on any hardware that the customer chooses, and then the software is able to manage that storage, so that's software defined storage. There are different flavors of software defined storage which we are here to talk about. One is hyperconverged infrastructure which VMware is pioneering with its vSAN technology, and we also have the Dell EMC ScaleIO product that is offering customers the capability to run multiple workloads from multiple hypervisors. Jason? >> Yeah, definitely, so you think about the evolution of the data center with software defined. It started with Compute. VMware did it back in 2001 with virtualization where they were able to abstract, pool, and automate Compute. Then we've got software defined networking which has been hot. Now the last part of the SDDC is storage piece. So the key message with software defined storage is to essentially abstract the hardware, the storage inside the hardware, away from the typical array type architecture instead running on off the shelf X86 servers instead. So you're simplifying the Data Center, and you're standardizing as well on more common componentry to reduce cost, to increase efficiencies, and things like that. >> Isn't this what SANS was supposed to be in the first place? >> Well, right. So weren't SANS created to get rid of DAS? And what we've seen is that customers are coming to us saying I have a SAN model, but I'm finding it to be a little too rigid. Or I'm managing multiple silos of SAN, and it's becoming very cumbersome. And then, hey guess what, what happens after three to five years? Data migrations, tech refreshes, dirty words, right? And so customers are facing some of these challenges, and so they're looking at software defined storage to help solve some of those challenges they're facing with traditional infrastructure like arrays. >> What you're doing is essentially you're de-emphasizing the hardware as a vital component in the equation. Isn't that something that Dell EMC wouldn't want to do? I mean, why would you want to make hardware essentially something you don't think about? >> Well, it's essentially a portfolio approach. So think about mainframes. How long have they been around, right? And they won't be going away. Same with traditional infrastructure. For certain workloads, perfectly acceptable to continue running them on traditional array infrastructure. What we're seeing with software defined storage is that for some of these third platform, next generation applications, as well as some less critical workloads, customers are looking to optimize the storage that is being run on it. Software defined storage does that by giving you common hardware with all the goodness, the reliability, the Enterprise great feature function set, the scalability, the performance currently lives in the software itself, and therefore just gives you a lot simpler data center to not only manage but also build. >> So you say SDS right now is not something that you would use in a mission critical application. It's something that has a lower performance threshold perhaps? >> So I would say it's not lower performance. It's applicable to most workloads. So if you look at the evolution of the market, X86 workloads have bigger mainstream. Virtualization is mainstream. Customers have successfully virtualized their data centers. Most of the customers that we engage with at VMware, have successfully virtualized their data center. For those customers that are there in their virtualization journey and have completed the virtualization journey, they're looking at how do we extend that virtualization to storage and managed storage within the same framework of a server without having to use a specialized storage array. So we are increasingly seeing that, and most of the customers that are looking at digital transformation and transformation of their data center are looking at how do we take advantage of the software defined storage. A majority of the applications are running on X86. There are a specialized set of applications that would still benefit from storage arrays, and they're going to preserve those storage arrays for those set of applications. >> Now in storage you typically would have tier one storage or your high speed storage, tier two sort of your mid range, and then your archival storage. Would you have different SDS solutions for each of these, or would you abstract across the whole range of platforms? >> Dell EMC's approach is a portfolio approach. So we're going to have a jack of all trades with it. So there's various products within the set. So for block storage workloads you'd have ScaleIO. For file based workloads, something like Isilon. Something with archive and object storage, ECS. So it's more of a portfolio approach in that sense. >> But from a customer's perspective, am I abstracting, am I software defining all of those at once in one common pool, or to what level of control do I still have with where that data resides? >> So you do have a level of control as to where that data resides. What we have seen customers do is take a step by step approach to software defining their storage portfolio. For example, with the VMware vSAN technology, we see customers that are using virtualization solutions today with the vSphere product that we have. They have excellent storage systems, and they're looking at how do we take a set of workloads that are running on external storage systems, bring it in to the software defined storage approach. So, as Jason pointed out, I think the portfolio Dell EMC, Dell Technologies portfolio, has all the assets for our customers to do this piece by piece journey. It's an evolution. It's not going to happen overnight. We believe it's an evolution, and customers will slowly start embracing software defined storage. >> So let's talk about use cases. Right now, what are some of the most appropriate use cases that you're seeing for customers adopting this technology? >> So we embrace vSAN technology which is hyperconversion infrastructure. The first design goal that we started the product with was VDI. We see a lot of customers that are using vSAN for VDI. Now we are at the sixth generation of the product, and customers have started moving Enterprise class applications onto the vSAN platform. They now run Oracle. We are certified for SAP. We offer all the Enterprise class features such as deduplication of another work, all flash platform, as well as erasure coding. So customers now are embracing vSAN for Enterprise class applications, tier one applications, such as Oracle, SAP, and mission critical applications. >> We see some of the similarities as well with ScaleIO. We target more Enterprise service provider businesses looking to consolidate block storage workloads, looking to build a private cloud, and looking to do third platform devops types use cases as well. So containers, open staff, things like that, so that's where we kind of target our use cases for ScaleIO. >> These are not high performance applications. They may be mission critical, but they're not high performance necessarily. >> They are high performance in that the iops requirements expected from the storage system are enormous. So for example, vSAN all flash, a single ready node of vSAN can provide about a hundred thousand iops, and we can scale up to 64 nodes. We are talking about millions of iops that are available by the storage sub-system, and we are able to provide a policy based management framework where they can apply policies with respect to what type of performance requirements the application needs and serve it out of the platform. So these are applications that have high performance requirements as well as a blend of mixed workloads that can be supported by the platform. >> Do you work at cross purposes at all with your colleagues on the hardware side where they're trying to sell the virtues of their specific hardware platform, and you're saying, "We'll put it all on software "and we'll all manage it as a pool." >> So we do leverage the underlying capabilities of the hardware. For example, NVMe is one such example of a hardware technology that is readily available from our hardware partners. We work very closely with those hardware partners to make sure that the software leverages the performance characteristics that the underlying hardware provides. So at VMware, we take this ecosystem approach where we work closely with the partners, make sure that the drive is capable of providing those iops, and we are able to take advantage of those drive characteristics, and provide it for our customers. >> This is the magic of software defined storage, right? Depending on your workload, and what you're trying to achieve, software defined storage can support it. So if you do have a high iops requirement, you got a bunch of servers with NVMe flash in there, put some 10 gig ethernet on the back end, and bam, you've got a high performance system powered by software defined storage, but if you don't have that requirement, then you could build a hybrid type environment where you've got some SSD's as a cache, but you've got more spinning disc in the back end. So that the magic of software defined storage that you're able to actually tailor the environment based on the workload. >> Whole new world out there. Jason Brown, Narasimha Krishnakumar, thank you so much for joining us, telling us about this topic that I knew very little about, but which clearly is going to be on the radar going forward. We'll be back from Dell EMC World Las Vegas in just a moment. This is The Cube. (electronic music)