 Live from Boston, Massachusetts, it's theCUBE, covering OpenStack Summit 2017, brought to you by the OpenStack Foundation, Red Hat, and additional ecosystem support. Hi, I'm Stu Miniman, here with my co-host, John Troyer. Happy to welcome to the program, Carlos Carrera, who's a Senior Principal Product Manager with Veritas. Carlos, great to see you. Yeah, thank you very much. Glad to be here. So, so many of the things we talked to here in OpenStack and the cloud world is relatively short-lived. The average lifetime of the average cloud deployment is like 1.7 years. You've been at Veritas a little bit longer with that, had an opportunity to have a conversation with you about some of your history. So, we're going to have to take the abbreviated format of that, but give us a little bit about your time at Veritas, some of the ebbs and flows of your career. Yeah, good. Well, again, thank you for having me here. It's great. We've been 16 years with Veritas. As I've mentioned before to you, back in 1994, 1995, we created the first file system on volume manager. A lot of things happened since then. At that point in time, the solidifying story stir was not yet there, but many years ago, we got some piece of software running on top of any kind of hardware and we was able to help customers to move workloads from one place to another. In a very agnostic point of view, right? And then we move into clouds and now, three years ago, we started looking into what do we do with OpenStack clouds? Because this is going to define, it's going to need something very new, something different. So, today, this week, we're very happy because we finally announced hyperscale for OpenStack, which is a solidifying story solution that has been built for OpenStack clouds. Yeah, when I look at the industry these days, the term lately is storage services, how we're doing things in software more. OpenStack is the open source infrastructure piece. You guys are the hipster player in this space. You were doing software-defined storage and software services not attached to everything else beforehand, so, sounds like OpenStack's a natural fit. Tell us a little bit more about how Verifah's fits into that. I think that, again, it was a perfect fit, but we had to review what we was doing, because, again, I've been many years, I was working with traditional legacy architectures in the past. We had our cluster file system that today can work with 128 nodes. But we revisit, is this what we really need to the new OpenStack clouds that are going to scale? As you said, what I need is the storage services. So, what do we have to do to provide those storage services to the OpenStack cloud? So, three years ago, we have this, we call OpenFlame project that today is hyper scale, has been really building from scratch. New provost is what we call emerging provost at Veritas. And finally, we got separated from Symantec and we got all the visibility on the storage again. And is using all the know-how that we have in the history, as I say, we're a very big startup, right? But now, emerging with new provost, with new solutions that has been designed for OpenStack from scratch. Can you drill down on the product itself? Is this file block object storage? Is this sitting on top of servers laid off in a server-based way? How does it interact with OpenStack drivers? That sort of thing. Yeah, that's a good question. So, it is syndera storage, what we provide is block storage for OpenStack. Something key, it is based on commodity hardware of your choice. So, you decide what is the hardware that you want to use. Really, X86 servers that you can choose in the market, whatever you want. And one of the key differentiators is that we provide block storage, but we separate the compute plane and the data plane. And this is an architectural decision we had to take three years ago, to say, we cannot scale, we cannot provide the storage services that you need in a single layer of storage. Because that is what most of the so-called defining storage solutions on the market are doing today. And then they're having problems with things like noisy neighbor, they're problems with things like scalability, like quality of service, and of course, they're having problems with protection. How do I protect my cloud environments with OpenStack? And we as a net backup company, we have our leading net backup solution, we hear that from our customers. But it is not about bringing another solution that is going to bring another noisy neighborhood. So we really have to separate two layers, compute plane where you have your first copy and the data plane where you use cheaper and deeper storage to keep the second, third copy and do all the data management operations. That's interesting what you just said there, like two copies. So you do have a copy that's close to the compute, but then you have another. Correct, because again, if you take a look to what you have in the market, typically it's one side fits all. So do you need three copies for everything? And today you have emerging technologies, you can have things like MySQL, or you need high performance, or you can have things like Cassandra, where you need nine copies at the end because the application itself is giving you the resiliency. So if you use a standard solution that for each open stack instance you have three copies, that means you have three copies, three copies, three copies and nine copies. And it's not only the number of copies, it's that when you make a right, you're writing nine times. And you're writing on the single layer. So we said, we have to separate that. The first thing is that, what is the workload? Stop thinking about the storage, stop thinking if this is a pool of SSDs and start thinking about the workload. And then we connected that very well with open stack because open stack you have the definition of flavors, right? That is how many CPUs do you need, how much memory, but also we extend those flavors to say, what do you need in terms of storage? What is the resiliency level that you need? What is the number of copies? What is the minimum performance that you need? What is the maximum performance? It's not only about solving the noisy neighbor with the maximum performance about limiting, it's about guaranteeing that you are going to have a minimum number of iOS per second. At the end what you can get, you can have a MySQL running with high performance needs with web servers on the same box without fighting each other. Carlos, can you speak a little bit about how customers consume this, how do they buy it, how's it priced, how do you get it to market? Because we've talked before with Veritas, storage used to always be either an appliance or an array or things like that and the software cloud world's a little bit differently, how does that fit? So today it's software only, so you make the decision about what hardware to use. We try to simplify the go-to-market model, so it's based on subscription. You just pay for the manage capacity that you have and you only pay for what you have at the compute plane. So I think a simple model that we could find to go in open source projects and being able to attach to that. Okay, can you speak to, when you talk about go-to-market from a partnership standpoint, it's a big market out there, Veritas, well-known name for many years, but what partners are involved in this, any certifications that are needed? So we're working with our typical partners that have some expertise with OpenStack and helping with them. We are now also working with hardware providers, we're working with Supermicro and creating reference architectures with them, because at the end, we have to explain to the customers what they can get from different hardware, so we're working with them. And we are also working with new partners, for example, yesterday with As on the stage, we have Furbanks. Furbanks is an open-stack ambassador in Netherlands. They have been working with As from the very beginning on this project, on the validation, and they understand OpenStack, they understand the issues, and they have been doing all the validation with As about, yes guys, this is the right thing you have to do from the very beginning. Is this product tuned specifically for OpenStack, or will it be available for other kind of private cloud? We have available for OpenStack, we're going to have it, we announced, I think we watched with you also guys, we announced the beta version for containers. At the end, the same thing is how do you provide persisting storage for containers. 90% of the product is all the same, is that compute playing, is the data playing, how can I protect my workloads from the data playing? Because again, it doesn't matter if it's container, if it's open-stack, when I have to protect it, how do I do it? How can I read my data without affecting the performance? And that's where we have the value with the data playing. And of course, our integration with NetBackup, our leader backup solution in the market, where yes, with a single click, I'm going to connect OpenStack with NetBackup, define how my workloads are going to be protected, when and how. Here at the show, OpenStack Summit, how has it been working with the community? Sometimes in the open-source world, vendors have to have a certain kind of conversation with that open-source community, then to show that they understand their needs and what they need out of the relationship. How has the week been then? So yeah, that's a very good question and let's go to something that we want to announce, hopefully at the end of the year. The first version that we announced this week is based on canonical Ubuntu OpenStack. At the end of the year, we are going to have a Red Hat and in our DNA is to be agnostic to the past any hardware and of course now is any kind of OpenStack distribution. So we will work with any of them and something that we want to announce at the end of the year is to have a community edition for HyperSQL. So again, that is our offering to the community so they can go use and provide feedback for us. And would that community edition itself be open-source or just available for the community? It would be available for the community. We keep our IP. As we get towards the end of the event, I'm sure you've had plenty of interesting customer conversations. Anyone, I'm sure you can't mention names but any interesting anecdote or just general feel of the community? My anecdote for yesterday when I had our presentation, we had a customer on the run. We have been working on a POC with them. We have been very, very helpful customer within us. Do you have any question? This guy's done up and went to the microphone and I was thinking, what is he's going to ask? He know everything about the product and he just said, hey guys, we are doing the right thing. This is great. I'm fantastic. You are bringing a lot of value here. So I was like, wow. And my understanding, it was a big brand name customer who actually said where he was from which is great validation. Something we've heard all week is there's that sharing here with the community. So financial companies who in the past wouldn't have done that. Telecoses who do that in the past. Great to see. Give you the final word, Carlos. Yeah, the thing again is, as you said, validation is a key thing. Being a lot of years in the company, I got this project eight months ago and all the things I've been doing is just validation, talking to customers, I don't know how many analysts I've been talking during this week and all of them said, yeah, guys, you are doing the right thing. This is the direction that we have to move. So happy that, you know, finally emerging again from where it has been back here in the community and up on the stack. Well, the speed of change, constant learning on new things and helping customers move forward. Big theme we've seen in the show. Carlos Carrera, appreciate you joining us here for John and Stu. Thanks for watching theCUBE here at OpenSack Summit.