 Live from San Francisco, it's theCUBE, covering DockerCon 18. Brought to you by Docker and its ecosystem partners. Welcome back to theCUBE. We are live at DockerCon 2018 in San Francisco on a beautiful day. Lisa Martin with John Troyer. We're very pleased to welcome back to theCUBE, distinguished alumni, Avinash Lakshman, the CEO and founder of Hedvig. Welcome back. Thank you, great to be here. So talk to us about Hedvig, what's new? What are you guys doing? What's exciting? Oh, a lot of things. I mean, since the last time we spoke here, I think some of the improvements that we have made in the platform has been pretty significant. In fact, we are executing on the vision we had from day one, which is to be the infrastructure for both primary and secondary storage. And I think we are delivering on that promise and doing so pretty efficiently. Talk to us about the case for consolidating primary and secondary workloads on one platform. So everybody understands how CAPEX can be reduced, right? But I think the key to understanding how one can reduce their operational overhead is going to become very important. And the key to that is to, if you look at how enterprises have evolved, there's always been a vendor they talk to for their sand needs, a different kind of vendor for their NAS. They're all trying to figure out what their object strategy ought to be. And now there is a lot of disruption happening in the secondary market. We always felt like, what if you could have one platform on which you can consolidate block, file and object, and also be the backup target in most of the secondary use cases. Then all you got to do is to train your folks on one platform and just use it for different workloads by changing either policy or hardware skew. Makes your operational overhead very streamlined and very efficient. I mean sometimes the infrastructure people joke about containers and they say, well, the developers here, they just go, well, the storage just exists, doesn't it? I just call some, I just call, I do a mount point and it just works in the cloud, right? And so, but I have figures here at DockerCon. Can you talk a little bit about some of the relationship there and some of the, how you work with folks, these folks here at DockerCon working on containers? See, anytime anything is simple to use, they always joke about it, but there's a lot of work that goes behind driving that simplicity, you know? Like they say, you want to keep things simple because any fool can complicate things, you know? So, but I think people are looking to bring that cloud-like mentality into on-prem data centers and I think we are also delivering on that promise. Yeah, and to, well, to just to clarify, yeah, I was kind of making a joke, the infrastructure people always kind of, they roll their eyes that they say, we do all this work and the developers think it's so easy once it gets to that. No, I know, because I mean, all the hard work is done for them. Exactly. I mean, there's also a joke among the infrastructural community that it's just an app. That's how they flippantly brush off applications. So, it goes both ways, I suppose. Indeed. So, I always look at messaging on websites of gas and companies and I liked what I saw. Finally, a cloud-agnostic storage solution. What does that mean from Hedvig's perspective and how does that give you that differentiation that you want? Good question, because for that, one needs to define what multi-cloud is, right? Everyone has their own definition for multi-cloud. Just the way a few years ago, everybody had their own definition of distributed systems, right? I mean, there was a large populace that believed that if a program runs on multiple computers, it's a distributed system. Not the way I would define it. It has to be multiple machines working together to make believe that there's only one machine behind the scenes, right? And similarly, one may want to set the stage for what multi-cloud is all about. Again, running, there is a popular school of thought if I run, if I can run on cloud A and then run on cloud B, I'm multi-cloud. That's not my definition. The way I would like to define it is one fabric that can span multiple clouds and give you the illusion that that's where this whole location transparency things comes in. You believe everything is local, but it could be anywhere. That's where infrastructure becomes kind of invisible and it's a single fabric that spans multiple cloud environments. So that transparency, is that something that really kind of helps Hedvig define some of your key differentiation? In the cloud environment, definitely yes. And here's why. When people typically build applications, if you program to services that are available in one particular cloud environment, tomorrow you want to run it on a different cloud environment, there is no API compatibility between different clouds. So you will have to rebuild your applications to a totally different set of services that another cloud vendor provides. At least what we bring to the table is from a data management infrastructure layer, you could have one fabric, you can now move your applications willy-nilly because you will be programming against an API that we provide and you don't have to worry about where you're running them. And that's what we enable. And that's done seamlessly today. Can you talk about how some of your customers are using Hedvig and some of the customer use cases in production? Yeah, I mean, there's quite a few use cases. I think the popular ones are for those who are in the primary kind of workload space, which are typically used when we need to run, when you need fault tolerance across multiple sites. One of the more interesting use cases that we are now deploying in the UK is across multiple clouds and on-premise. So it's kind of hybrid and multi-cloud built inside one fabric. And the reason they do it is basically not only consolidate their existing on-prem data centers, but also to satisfy data governance laws for certain applications. And this leads to a term that I'm trying to drive and make popular, which is declarative data sovereignty, which basically means if you look at things in Europe for different applications that are different laws, and that's now becoming kind of commonplace. We live in a global economy, but the data governance laws are all local. So people want certain applications and its data not to span certain geographic regions, right? How do you make that happen in a declarative way? And you should be able to do that by just saying for this app, this volume, and there are some policies you assign to it and that policy basically dictates what regions that data will live in. And we make it as simple as it possibly can get. So the policies then drive where the data lives and it will never, on head big it won't, if it needs to be in stay in Germany, that volume will stay there and won't ever be in. Exactly, and you can even control that across different cloud vendors, right? If you're running, say, for example, this particular customer, I'm not sure if I can mention the name, but so I'm going to err on the side of caution. They run across AWS, Azure and on-premise and for certain apps, they want the data to be spread across AWS and Azure. And for certain apps, they want the data to be spread across on-prem and AWS. You want to make it as simple as possible, make it declarative, you know? And for those who are more system savvy, if you look at how computer science has evolved over time, if you look at the late 90s, transactions were a big deal, right? Everyone were trying to figure out how to program transactions into their system. But over time, people designed runtimes where you can declaratively annotate sections of your modules to say whether it should be part of a transaction or not, and they made it that simple. And the runtime kind of takes care of that, delivering those annotations that you declare. We want to bring the same simplicity for data sovereignty, right? So we're at the fifth DockerCon and this morning, Steve Singh said it was around 5,000 and we've heard upwards of 6,000 attendees here. And I think he said at the first DockerCon there was only about 300 people. I noticed when I walked out of the general session this morning and I turned back around the room, it was packed, it was standing here mulling. I'm curious what your thoughts are about some of the things that Docker has announced. You know, they really talked about what they're enabling with enhancements to Enterprise Edition, with federated application management, what they're doing with Docker Desktop as really enabling three things. And I heard you kind of talked to one of them for sure, which was agility, choice and security. From a security perspective, what are some of the things that Hedvig can enable your customers to achieve? Because we hear security is a huge issue. Security typically manifests itself in two ways. But one is in today's systems you have data spreading all over the place. So when the data is on the wire, you need to encrypt things on the wire so that nobody can sniff and steal data. There is also an aspect of security where data which goes and resides on any media needs to be secure, meaning it needs to be encrypted. Now, there's a whole school of people who believe that they have solved that problem by sticking in what is called self-encrypting drives. But that solves only part of the problem and it's a hacky solution because we live in a world where BYOD was a big thing and now it's kind of BYOK, bring your own keys so we can encrypt using the keys that you bring to the table and get out of the way. So if you want to achieve that, then you cannot just use SEDs. You have to drive the encryption onto the media on which the data is going to reside. But that's part of the problem. The other part of the problem is what happens to data when it's on the wire? So you want to be able to encrypt data on the wire and at rest. So that takes care of one part of security. The other big issue is ransomware. And there are also application developers and they want to churn through their features and they want to build fast and deliver features very fast. They could land up corrupting data underneath. How do you protect against? You want to have a feature that even lets you protect your data from yourself. And those are all capabilities that one needs to think through. The way you typically would do that is have the capability of providing managed snapshots where you can periodically keep taking snapshots of your data so you can revert back to any. It's kind of like GitHub for data. That's how we look at it. When you look at GitHub, you put your source code in there but nobody is working on the same source tree. You create different trees and then you merge them in when you think the time is right. So you want to have different copies of your data without having copies and then be able to revert back to any pristine version that you deem fit if you make any mistakes at the application level or if there are ransomware kind of issues. So it's a multi-fold problem. So you got to look at things holistically and make sure that everything is kind of delivered natively through any infrastructure platform one is building. So Aminash, you are, well frankly you've had a big impact on cloud computing. You were one of the co-inventors of DynamoDB at Amazon? Dynamo. Dynamo? Yes. And you were, and then at Facebook Cassandra and you actually said the 10th anniversary of Cassandra. Yeah, apparently this is. I didn't even realize. I think it's in either this month, I think it's this month or next month. Nice. Nice. Huge impact, right? Whole ecosystems have been built up. So you have a sense of where the problems are. Now, Hedvig, a few years old now, you know what are you trying to do now with Hedvig and where would you like to take it? You're in the middle of a very hot market the data services market, this kind of secondary storage market is super hot. So I would just love to kind of hear what your dreams are, what you're trying to take Hedvig and what you would, you know, how did you see that this was a hot place and how are you going with it? Dreams are a difficult question to answer. At least we know one thing. What we do know is we are onto something big and this is a problem that needs solving and it needs solving from the ground up and it needs solving in a very different way. And I always believe that true innovation comes from people from the outside. I'm not a storage guy, I never was. But I believe that makes me better suited to go after this problem because I don't have any of the baggage that people typically tend to have, you know, when they try, you can't disrupt yourself, right? So we believe we are onto something big and so we are just heads down, trying to deliver on that. I don't know where it could go. I'll leave it to destiny. But I think we are onto something big and hopefully we will reap the benefits of what we are doing, right? Great. Well, Avinash, thanks so much for coming by theCUBE again and sharing with us what's going on ahead. But it sounds like some exciting times ahead and we wish you the best of luck. Thank you, really appreciate it. We want to thank you for watching theCUBE. For John Troyer, I'm Lisa Martin. We are live from DockerCon 2018 in San Francisco. Thanks for watching guys.