 Live from Seattle, Washington. It's theCUBE, covering DockerCon 2016. Brought to you by Docker. Now, here are your hosts, John Furrier and Brian Graceley. Okay, welcome back everyone. We are here live in Seattle, Washington for DockerCon 2016. It's the SiliconANGLE Media's theCUBE. It's our flagship program. We go out to the events and extract the signal from the noise. I'm John Furrier, my co-host Brian Graceley, another hot startup up here. Robinsystems, we've got the CEO and CTO, CEO, Pramal Bhuch and Partha Satala, CTO of Robinsystems. Welcome to theCUBE. Thank you. Good to see you guys again. Again, you guys are startups. We've got 15 million in funding last fall. You're doing something a little bit different with storage. Unshackling applications that have a container platform. So, explain that. It seems to be a lot of people focusing on storage to make storage go faster, better, all flash, what we see the whole time. But you guys take a different approach. Explain, take a minute, explain what you guys do specifically and why it's not just another storage container startup. So, what we are doing really is looking at how to apply containers to the problem of virtualization. So, what is a piece of it? But it's not everything. Really, when we look at applications like databases, applications that consume a lot of data, big data, Hadoop, Spark, even no SQL data like Mongo, Cassandra. What we see is that these are traditionally run on bare metal. Virtualization hasn't really made a lot of headway because they're performance issues with the hypervisor performance tax. Containers are a great solution. They eliminate a lot of the hypervisor performance tax, but they haven't really addressed the issue of data gravity. When you've got terabytes of data, how do you really make that portable so that you can get all the applications virtualized on containers? See that storage startup? You're a software startup. Exactly, and so, we believe that you need to look at the whole virtualization problem end to end. If you really want to truly solve the performance issues, provide performance isolation when you pack multiple applications on the same physical hardware and do that in the context of containers. Yeah, so it's a very different approach. You've got a lot of things going. I've got software defined storage. People are still working through. You've got containers, which a lot of times is very kind of new application centric. We're seeing this movement back to people saying, hey wait, it works for enterprise applications as well. Walk us through how you guys think about performance when you're dealing with software, you're not dealing with hardware anymore. Help customers understand that mindset because that hasn't necessarily been the case with storage and data management in the past. It's been very hardware centric. Yeah, go ahead. Yeah, the big difference between a traditional app that's continuous today and the enterprise apps is that the enterprise app is more complex and its need from the storage side is more complex. But beyond that, once an enterprise app is deployed, there's this life cycle management that becomes very important. Things like snapshotting, things like cloning. Or when it's done in the context of storage, just a storage subsystem, that problem has been solved. Once you elevate that to an application construct, that is more complicated, especially with the modern day distributed applications. You're not talking about a single volume that's being snapshotted, but you're essentially looking at a collection of volumes that spans the entire topology of an application. And in fact, spans the entire service that an application depends on, whether it's a database, an application server, and so on. The entire life cycle has to be done from that context. And again, from the performance point of view also, if you look at a complex service which has databases on the bottom and an application service on the top, each of them might have a different performance need in terms of an IO ops or throughput and so on. So essentially you've got to enforce the IO ops or the throughput upper and the lower limits based on what the application needs. So that essentially changes once you look, I mean, essentially infrastructure management changes once you look at it from the applications context. So all the big folks, we had Jerry Chen from Greylock on who always likes to come on the queue. He's been a CUBE alumni since we started and he's always got his finger on the pulse. I asked him the question, where's the focus? Let me use all my chat box and all this stuff. And what he pointed out was really kind of what you guys are having core focus. That's the applications are driving the infrastructure policy stack characteristics. One, do you agree with that? And two, how do you talk to customers and say, hey, I want the infrastructure just to work? I have VMware, I got virtualization, I want to virtualize, but I want my app workload to be really awesome and fast and guaranteed. I mean, what's the conversation like? Take us through that. So I mean, so far we've been looking at either infrastructure and isolation of apps, right? You're looking at the storage, you're looking at just containers by themselves and so on, right? So the conversation that you have with customers is essentially what can we do for their applications? How can essentially move application awareness into the infrastructure, right? And that essentially drives a bunch of use cases that you would not really look at here. A pure play, let's say a storage for containers or a pure play orchestration, for example, right? You want to add something more? Is that their pain point orchestration? I mean, where's the aha moment for your customers, for you guys, when you generally walk in there? Aha moment, I would say is predictability, right? Performance predictability, at the same time, given the consolidation, right? And that is something that you cannot achieve today, right? The moment you essentially put multiple apps in the same host, whether it's containerized or not, predictability is gone, right? In fact, some of the demos that you're showing at our booth, that customers come and ask us, how do you guys do that, right? How do you essentially enforce both max IOPS, min IOPS, and how do you do it across multiple apps? So what do they focus on when you guys give them the demo, when they ask those questions, obviously they're impressed, they're trying to dig into what the value is, and when you're at secret sauce, but what gets their attention, what's the one thing that in the demo that you guys are showing that gets the customer's attention? How do you do it, I guess? How do you enforce? Yeah, it's magic. So you talked about big data, you talked about Hadoop, you talked about Spark, you know, some of these kind of fast-moving things, a big deal for business, because it drives better intelligence. You know, infrastructure's always been a little more, it's a little slower moving, it's resilient for a reason. I mean, is what you guys are doing really trying to disrupt that balance that says if the applications are going that much faster, it's got to be software, it's got to be delivered performance, but I need to make it simple to deploy, simple to move things. Is that really the, you know, John talked about the AHA moment, but is that really the thing that people are looking for a new balance, a new equation for, to make that work? So a lot of people we are talking to, they are trying to figure out what all these new big data applications mean for infrastructure. And in many cases, they are kind of kicking tires trying to figure out what they do with containers, infrastructure, storage requirements, what the application is going to be like tomorrow, and how do we future proof the infrastructure. And I think that's a conversation where we are essentially an infrastructure layer, application agnostic, but we provide them that whole future proof portability, that's what they get some interest in. And are you guys, do you have the capability of somebody who says, I want to start small on top of Amazon or Azure, but also have an on-premises deployment? Can you link the two together? Are there ways to take advantage of different sort of resources? Yeah, so today most of our customers are on-premise customers, but we have several conversations going on on the hybrid cloud aspect where they're trying to move some of their workload into the cloud, and how they can do it with a single pane of management, with Robin being sort of the common layer that enables them to do that. So you guys talk about the application centric data center or application defined, software defines pretty much what's been the big buzzer, application specifically, what on the product side, how do you explain that to customers and say, what does that mean to, first, what is application defined data center mean to you guys? So if you look at virtualization, the traditional virtualization gives you VMs as a proxy to physical machines. When it comes to deploying applications, you still need to coordinate between the VMs to get the application up and running. What we are saying is that on top of providing that entire virtualization stack, compute network storage, we're also adding an extra layer of abstraction so that the customer, the user only needs to worry about deploying an application at a level of, give me 10 node cluster, and we are doing all the plumbing beneath that. So the whole combining that image management, deployment, orchestration, along with the traditional compute network storage virtualization, that when you put it all together, that's what we really talk, I mean when you say application defined data center. So Farth, I got to ask you the product question. You guys are startups, obviously you're pedaling as fast as you can. You got customers, you got some big names on the case studies you're working with. What's the product focus right now as priorities for engineering? What's the, and obviously the basic, it's stuff out solid shipping, but what new things you guys working on? What's the roadmap look like for Robin systems? So I mean we started off with focusing on data applications and at least in the near term we won't essentially focus on that and have a more complete story so that we essentially span all types of apps. The last cube we did at Strata, we showed a demo of how we do a Hadoop and how customers are running Hadoop, right? We're expanding that to a lot of other applications like Oracle, Cassandra, Mongo, modern day as well as the traditional enterprise databases and so on, right? And then the next step for us essentially would be to essentially expand this from just an on-prem solution into the cloud. And when you say cloud, it's not just taking cloud and then deploying our software there, but essentially taking the characteristics of a cloud, it's instant expandability and the elasticity that you get with cloud and then weave that into our product so that the customer gets a very seamless experience deploying these apps. So I'll compare and contrast the Hadoop ecosystem to the Docker ecosystem. We love to talk about the Hadoop because we've been covering it for a long time. We've seen the evolution and our point at the last Strata was we're still seeing customers saying it's still complex, I mean it's getting better certainly, but it's still complex and the total cost of ownership is high, but it's getting better. It's certainly working, big data, no brainer. Docker, we're trying to tease out the same thing. It's complex in the sense there's a lot of noise here, people who do different parts in the stack. How would you guys compare the two ecosystems? Because Docker seems a little bit more agile in terms of less cost, because it's very developer focused. We're trying to put our finger on cost of ownership and can you guys comment about your thoughts on the two ecosystems? Okay, so since you brought up two examples that have Docker and Hadoop, right? So I think the way you look at the problem is very different. I think the way you look at the problem from a Docker angle is a very container-specific thing. It's not really application-specific yet, and I'll expand a little bit on what that means. But when you look at a more complex application, it's not just one container. It's not even just a collection of containers. It's basically a group of containers playing different roles, and all of them are interconnected with each other, and each of them depends on the other for different things. And when we deploy these things and we essentially want the customer to look at it not from a container or like Trimble said or the virtual machine, or look at it from a pure application perspective. Just to take an example, right? Let's say you're deploying Cassandra. You would go and say, I want to deploy a Cassandra cluster. You don't go and say, I want to have these many VMs. You say, I need Cassandra. It requires this amount of scalability, so I want to create those many number of partitions. Each of those partitions requires different IOPS. How do I set it up, right? So essentially, define your application needs from the context of an application. So move at a speeds and fees just getting, standing up the resource, basically. And Hadoop has had some. So Hadoop is actually a very big example of that, right? Because Hadoop is a very complex thing. Just to get a Hadoop cluster going is very difficult. And in our product, it's essentially a push of a button. You define, go to the application creation page and just click. Easy button, as they would say. Easy button. One click, one click deploy of any complex data. Okay, so I got to ask the question on the website. You say unshackle your apps. What does that mean? When you say unshackle your applications. What do you guys do to unshackle apps? So let's say you are deploying a virtual machine today. You as a developer are using a virtual machine. Very rarely do you know where it is running. All you know is you go to the virtual machine or the VDI, let's say admin, and you say, I want to get a VDI thing, right? You don't really care about where it's being deployed. We essentially want developers to look at application deployment in the same way. Forget about asking for first. You say, I want to deploy this app, right? So once you do that, you're essentially taking away applications from being dependent on infrastructure to just defining what you want and you just get that magical thing. So the whole composable infrastructure kind of narrative that's been kicking around, certainly HP's writing that happy, is what you guys are offering. Exactly, and particularly for the applications that we are going after, they tend to be run on bare metal silos. So when you talk about unshackling, it is about that fluidity, the portability that we are talking about, where you don't really need to do physical provisioning for specific machines. We can completely take away the deployment times, the provisioning times, and make it all portable. That's the key part of the unshackling. This is a real changeover. Guys, I'd like you guys to spend a minute, talk to the audience about Robin Systems. Take a minute to share. There could be potential customers out there watching. Why Robin Systems? Why should they be engaging with you guys? I think today the biggest problem that we see with data heavy applications is that there is no real solution that allows the high performance and performance isolation that you get with bare metal silos. With the flexibility, the agility, the portability, with things like containers and VMs, we are trying to bridge that gap. And I think what you get by working with Robin Systems is a seamless fluid infrastructure where you can maximize your hardware utilization but also make it very easy to manage and deploy applications. So you have big focus on the delightfully easy to use interface, where you take away all the complexity of the plumbing out so that you can just focus on what matters to you, which is application management. And we take out of all the headaches, we're managing infrastructure behind that. Taking the complexity out of DevOps, taking all that complexity out of the Ops side, building the software. Like you said, I want Hadoop, just give me Hadoop. I want Cassandra, just give me Cassandra. So we've talked about that in our research. That's the way it should work. Yeah, we've talked about that from a research perspective. We think there's going to be about $300 billion out of that big cloud market that's just going to get embedded back into software and take it out of operations costs. Yeah, it reminds me of the famous Steve Jobs when he took over Apple. He moved away from speeds and feeds to solutions and what are we trying to do kind of thing. So that's awesome. Well, final question I want to ask you guys, and I'll put a little different spin on the question is, what's the coolest thing that you've seen your technology be used for in a customer scenario? I think for me, the most exciting thing has been how one of our customers who was really struggling to figure out whether to go on cloud, whether to stay on bare metal, their decilos, they looked at our solution and shut down both of their cloud and the bare metal plans. And we are going to use you to virtualize the entire deployment and that, when that aha moment happened to them and they changed their entire strategy and mood, completely locks up and barrel behind a Robin based solution, that was very exciting. Selfishly as a startup, it's a great first moment validation. The first customer really gets it, right? Totally gets it. And they actually change their infrastructure over. Absolutely. Do you think you agree at your coolest moment? Absolutely. What's the coolest technology that you have in the secret sauce tool chest that customers go crazy for? Can you share any tidbits on the IP? Well, from a technology perspective, I have not seen any company essentially bring applications into infrastructure, essentially making infrastructure more application aware, right? It actually spans a bunch of things from a technology perspective. It's not just not storage, it's also thinking about lifecycle management. So I think that is the core IP that we have. I've not seen anyone do that. Yes, it's a winning formula, application aware, fluid infrastructure, pools of resource. Robin systems here on theCUBE guys, thanks so much, co-found hot start up here at DockerCon 2016, it's theCUBE. I'm John Furrier with Brian Gracie. We'll be right back here watching theCUBE.