 Hi, Jeff Rick here in the Powell offices of Silicon Angles theCUBE. We're here for a CUBE conversation. We like to do these every now and then, get updates from some of our friends, find out what's going on, and we're really excited to have Noam Shandar in. The COO of Zadara Storage. Zadara is really doing well, and we just saw you most recently, I believe at AWS Summit in San Francisco. So good to see you again. You too. So I was just joking, doing a little bit of background before the interview here. There's no such thing as an overnight sensation. We first talked to you way back at AWS Summit San Francisco 2013, so it feels like suddenly these guys are just kicking tail, but you've been at this for a little while. Yes, yeah. Talk about how the market has changed, really, as your product development has changed and your maturity, now you're going to market, you're getting a lot of customers, but talk about the change of the attitudes of cloud and the attitude about AWS specifically, and how that really has morphed over, was amazingly two or three years. Yeah, so a couple of things have happened, and it's kind of interesting. So one is, yes, when we started, it was very difficult to convince customers to move into the cloud. So we had to go into the cloud and find who's already there and sell to them, whereas now it's not even a question. So we as an industry have cleared that hurdle. Nobody's asking why cloud or, should you know, is it safe enough? Is it good enough? Right. But the other thing that's super interesting is exactly a year ago, we announced our on-premises service, and that's flying off the shelves, which tells you that as awesome as cloud is, there's still a huge market on-premises. It could be things that aren't allowed to move into the cloud, or it could be things that just aren't ready yet. And at least as far as we see, it is still a far larger market. So we see, I think we're hedging, because we are in both places at the same time, customers can choose and they can change their mind later. So we don't really have a strong opinion on which is right. Right. But Pat Gelsinger and a whole bunch of people hybrid is the answer, right? There is no absolute answer. It's all application-specific. Application will drive the requirements for the infrastructure, as well as the storage is being part of that. So that's not inconsistent. But what's interesting about what you guys do, and maybe we should just let you give a quick kind of 411 on Zidara for people that don't know the company, is you still use the cloud as a method, as a way of doing business, as a way to deliver your services, even if it's not actually in a cloud, if it's on-prem. That's right. We observed when we started, but like you said, it's not overnight success we've been at it well over four years. We observed in the very beginning that customers like the cloud business model and it's not just the business model. It's the ease of use and the simplicity and the lack of headaches. So all of a sudden the burden of ownership is gone. I don't have to worry about maintenance. I don't have to worry about hot fixes or firmware upgrades or anything like that. It goes away and yet the product gets better over time just without downtime. So what we saw as customers were realizing, wait a second, when I use Amazon, life is really easy, why is this thing that's sitting in my data center so difficult? And we saw it and succeeded in making a product that is as robust as the traditional data center product but offered with a pure cloud model, with the live upgrades and are taking care of things behind the scenes so that customers don't have the burden of ownership and magically, but not magically through good design, they save money. So we usually think of renting a car or even having a chauffeur car is being very, very expensive but our value proposition is it doesn't have to be that way. You can actually have the chauffeur car experience for less than buying the car. Okay, so you covered a lot of things here. So first off is kind of the basic cloud mantra which is be elastic, use what you need, when you need, get more when you need it, shut down, make it go dark when you don't and you're delivering on-prem storage capacity with that business model, correct? That's right. You find people like that. People love it. Again, they obviously love the concept. Their biggest question is, well, what is it gonna cost me? Because they're thinking about the rental car which is more expensive in the long term than buying a car and we tell them what the price is and we also now have commissioned a study that shows you say, obviously save money on day one because you haven't purchased anything but the longer you own it, the more you save. And even after 10 years, you're still at a 50% savings versus the traditional storage. And we get it done two ways. One is as commodity hardware. So we don't have any component that isn't manufactured in at least tens of millions of units and some in hundreds of millions of units. The other is we remotely operate these, call them appliances and we designed these appliance to be easily remotely managed by us which means that the marginal cost for us to add one more is almost zero. Whereas traditionally when you buy a piece of storage and storage array, you have to dedicate a person to that. So we dedicate a small fraction of a person and that's a savings, a significant savings versus the total cost of ownership and we pass those savings on to the customers. And that's the part I guess that you would say or at least came out of your study that most people miss about the rent versus own because like you said, classic kind of thought process would be yes, I can rent but at some scale, owning is buyer, I can amortize my cost over all this business I'm doing and all the utilization but what gets left off is the operating expense that goes and wraps around the actual hardware. Yeah, absolutely. And by some studies, those operating expenses could be half again as much as the cost of the product every year. Right, right, and they don't get cheaper like storage tends to do every year. That's right. Yeah, so that's great. So let's shift gears a little bit. We cover a lot of hot topics. We go to a lot of cool shows. AWS was super new three years ago, not so much anymore. We're looking forward to going back to reinvent here shortly but one of the really hot spaces right now is Docker. And we took the cube to DockerCon and it's all about containers and it's just kind of this next manifestation of the consumerization of IT in the expected way that apps are supposed to work and a big part of that theme is like you said, remove all the headache. My developers just want to get to work and build cool apps and deploy cool apps. You guys have some stuff going on with Docker. So I wonder if you can give us an update on what Zadar is doing with Docker specifically in containers in general. Absolutely, so in June, we launched what we call Zadar container services. And Zadar container services simply put is Docker running inside the storage. The beauty of that is that now the storage and the computer are in the same place. So if you need very low latency access to the data or very high throughput access to the data, this is one way to get it done. And it's using the same cloud business model as the storage. So you get what we call is ECS engines. ECS being the Zadar container services. And you can size that engine any way you see fit from very small to very large. The smallest one is free. So we want our customers to experiment with this because it's a new way of doing things. There's not a track record of running Docker inside storage. So we say, hey, give it a try at no cost. And then larger engines cost an increasing amount of money per hour. So you can use it for an hour or two or a day or a month or a year. It doesn't matter. We charge you based on the actual usage. You can dynamically grow and shrink it. So let's say they have a task that has a shifting need in terms of the amount of resources that it consumes. No problem. You can auto scale it up or down and you can shut it down when you're not using it. So just to be clear then, so what people are doing is basically writing their app as they would normally but then they're deploying it to a particular Docker container. That's the one that's sitting inside is at our storage. Exactly. And by running Docker with a full fledged storage system, you get some things that you don't get with traditional Docker because when Docker runs on the server, it gets access to the local storage but it doesn't get shared storage. It doesn't get persistent storage and it doesn't get high availability storage. So for those uses where you don't need those, fine. So there are certainly plenty of uses for Docker where you have an ephemeral data set and you don't need to keep it. These are sometimes called stateless applications but as soon as you're stateful, all of a sudden there's a question, how do I do that? And this is what we enable. So the data is persistent in its high availability and it's shared. So multiple Docker containers and also multiple Docker containers and multiple VMs can access the same storage at the same time. It's pretty crazy. This thing just keeps getting more and more virtualized, more and more virtualized. I wonder, so the other than kind of knock on Docker, there's some stuff on Twitter today is does it open up a new security hole? Some had a great, are you willing to bet your job on the security of the container? So how have you guys addressed that in terms of adding this new element into your kind of controlled system? Yeah, it's a great question and it's one of those things that we get because of our architecture. So the architect, we call our product the VPSA, the virtual private storage array and the P in VPSA is very important because the privacy is achieved through dedicated resources. Each customer gets their own VPSA, one or more VPSAs and each VPSA has dedicated CPUs, dedicated memory, dedicated drives, both SSDs and hard drives, and a dedicated network. So anything you do in the ZCS and this is our container services is inside your VPSA which means it's running on dedicated cores of dedicated memory inside your private network which means it's every bit as secure as whatever your network setup is and it doesn't open up any new holes because it lives within this bubble, if you will. Okay, okay. Do you have any customers that have put this to use to talk about? Yeah, a number of them and we've publicly talked about a company called StudioTac, they're out in North Carolina and they are in the architectural and construction space and they are a service provider to these companies that deal with a lot of drawings and plans and those come in many different formats and there needs to be a conversion from one format to other formats for anything that comes in and that used to be done in batch fascia with the server running these conversion processes and that's fine and good but there are sometimes as many as a million of these files and even one or two milliseconds of latency between the storage and the server add up to a lot of time. So this is where when we announced to this customer that we're doing Docker containers, they jumped for joy. They said this is exactly what we need. So they already used Docker so we didn't have to introduce them to Docker and they already containerized their conversion process and now all they need to do is upload this container into the storage and now the conversion is done both far, far faster because now you're not multiplying a millisecond by a million files but also it's automatic because it's event driven. The storage informs the container a new file has arrived, what would you like to do? Right, right, let's do that process, interesting. So I wanna shift gears a little bit about, a little bit, just real simple stuff. You guys list 100% SLA, right? You're beyond the nine, is that accurate? So if the drive goes down in this unit that's in my network, how do you guys, how's it get fixed? So the very short and glib answers, it's not your problem, it's our problem. So the system is over provision, it has a lot of spare drives and in the event that a drive fails, two things happen. One is rate protection kicks in so you still have access to your data no matter what and then the other thing is a substitution of a healthy drive for an unhealthy one and this can be done completely automatically and transparently in a matter of seconds or you can do it manually with three commands and you can script it or anything like that. Most of our customers actually choose to do this manually because they don't wanna pay for an extra drive that's on standby because it's just three API commands to do a drive replacement so they say I'll just do that. It'll take me a minute instead of a second but I'm happy with that. But the beauty of it is to you, you have what looks like an infinite pool of drives, you can replace drives anytime for any reason. Maybe you wanna change drive type or maybe a drive failed and it's up to us to periodically sweep through, take the bad drives out. For those customers who want it, we shred those drives which is kind of, if you ever watched it, it's really cool. It definitely- Like a big chipper? It's like a big chipper. For those of us with an appetite for destruction, it's really cool and for other customers, we just wipe them. I mean, in fact, the customers have the option to perform the wipe themselves that's built into the system which is good practice before you return a drive into circulation, you wanna make sure that it's clean. Right. I wanna delve into another piece of the puzzle I don't think a lot of people talk about and you've talked about it before. When people are comparing kind of CapEx versus Apex and they're talking about having these elastic systems that really move with your application demands. And you made a really interesting comment on a prior interview talking about planning cycles. Right, everyone kind of forgets about budget planning cycles except the sales guy when he's trying to make it close, right? He knows that budget cycle pretty well but your budget cycles don't necessarily work with your planning cycles which don't necessarily work with the application uptake your application delivery cycle especially today in a world of agile where people are pumping code out like crazy. I wonder if you could talk about some of the experiences that you have had with customers potentially that really see now the value of aligning not only the purchase and the gear with the apps but that painful process called the budgeting cycle and the procurement cycle which is real. Yes, so as you said it's painful regardless of timing because with the traditional CapEx model you're buying something very expensive and you're stuck with it for years upon years so you better choose the right thing which means you have to do a lot of research you have to forecast your own needs and then you have to compare your needs versus everything out there in the marketplace it takes many months and at the end you're wrong anyway because it's impossible to predict the future and also you're so afraid of being caught short that you over purchase. One of the things that showed up in the TCO study that we did, the total cost of ownership is on top of all the other savings I mentioned there's also the savings of not having to over provision you only pay for what you use you never pay for idle capacity so that's yet another efficiency that we introduce so very painful cycles and you repeat every few years it's literally a Sisyphian effort you roll the rock to the top of hill and then three to five years later it's back at the bottom of the hill and you start over but on top of that is what you mentioned is the budgeting cycles we have one of our customers is Cal Poly they're a university, they're part of the Cal State system they're dependent on state funding which comes only once a year if they have a need in the middle of the budgeting cycle too bad, you can't so when they discovered us they said this is it this is the solution, it's only OPEX I don't have any upfront investment and I can source it whenever I need it if I need more during the academic term I use more, if I need less right now they're off, it's summer guess what? their system is idle and they're paying a very small amount because they're not using it it's perfect so they save money and they adjust to their actual real-time needs rather than to something that they thought they need a couple years back well and in a world where you had your SAP installation and you didn't really swap it out that often you could at least have a little bit more accuracy on your forecasting on your demands because your transactional growth your storage growth all those things had a much more predictable pace but now things are just going bananas and you just don't know and the demand for apps is just going crazy and then oh yeah there's this whole big data thing going on in the background as well so what if you've been at it for a while you guys are in the marketplace you're having some success what's kind of next? where do you see kind of the next big wave big challenge, big opportunity both versus it are but kind of in general as this market continues to move along this cycle so let's start with the market in general I think we've now proven that as a service is what customers want we prove it not through market share because we're still we're on a very aggressive ramp but we're still small by the larger scheme of things but we've proven it by the kind of customers that we've brought on board and by the companies who didn't win the very same bids when I would also add and also by the applications in which you're supporting correct probably really important we're talking about mainstream applications including mission critical applications at companies up to and including Fortune 100 and Fortune 50 companies so I think we can now state with confidence this is for real and the challenge for the industry is to get with the program there are companies who are resisting as a service trend and it's understandable it requires retooling everything so it's a monumental task and then there are companies who are service washing so they say we're as a service I'm not going to name names but a reporter in our industry wrote a recent story about as a service offerings and when you read the story all the companies mentioned do not offer an as a service offering they may provide some capacity on demand or some flexible pricing model but they don't actually provide a service so parts of the industry still don't distinguish what is true as a service and what is not and we talked about it in one of the previous interviews to me that's the litmus test do you have the headache of ownership if you do it's not as a service right okay so that's everyone's accepted it where's it go next I think it's going I think it's going to sweep across and some companies will adopt it faster but where it goes next is IT in general not just storage but IT in general will be consumed using a cloud model because that's what companies need they need to and by the way it'll cascade very quickly if competitor X has the ability to be very agile deploy new applications quickly and serve their customers better and sooner then competitor Y has to get with that program or else they'll be left behind so what we do is exactly that we enable companies to decide what they want to do and then just do it whereas with the old model is here's what I want to do but here's what I can do and they're not the same because I've locked into an IT infrastructure that didn't anticipate what I need right now right but then there's also kind of that locked in application in the stuff that runs on that infrastructure that for a lot of it's probably easy just to let it be, let it run I always go back to like the classical the airline reservation system when you watch the agent behind the counter tap into that system there's no reason to change that out that thing works just fine and then what about for Czar what's kind of the next hill for you guys to climb our biggest challenge is scaling so we're growing very quickly so I became COO in February of this year we didn't have a COO until then and so we needed to reach the next level of maturity and have the structure in place to keep growing at the pace that we've been growing so when we look at it is we're in such a good position and the customers are so passionate about the product it's how do we keep adding customers at the same pace keep and honestly it's really on us because when we go into a customer and they like the model we win, there's not a good alternative for that so that means we need to visit as many customers as possible right, right we just have to have the queue but this is our summit I saw you guys were all chilling out and looked like a very cold place to have a drink with your jackets all right, we'll give you the last word before we let you go what can we look forward to I'm sure we'll see you again at Reinvent you have some exciting news that you're working on some new things coming out we have exciting news all the time but we're going to focus anything takes time to adopt so we're going to focus on Zadar container services we're going to focus on backup test 3 both very popular features and we will I think you'll hear us talking about those features and the adoption and actual customers using those more and more until we're ready for another feature announcement so we try not to move too quickly for the market because then we end up with features that aren't being used right, well passionate customers are always your best sales force and we see that time and time again at the shows that we go to like at ServiceNow it's blank in a lot of these places where they just have a parade of customers that are really excited about using that technology because it's really delivering value to them so that's a good thing to keep working on awesome, well I know I'm shut there thanks for stopping by from Zadar Storage you're watching Jeff Frick keep conversation from Palo Alto see you next time