 Live from New York, it's theCUBE. Covering theCUBE, New York City, 2018. Brought to you by SiliconANGLE Media and its ecosystem partners. I'm John Furrier, Dave Vellante. Been here nine years. Our nine years of coverage. Two days live in New York City and our next two guests, John Murasic, CEO, Manistated, Peter Schmelz, CMO of Mystic. Good to see you again. Welcome back. Welcome to theCUBE. Glad to be here, guys. So obviously this show, we've been here nine years. We were the first original Hadoop world. We've seen it change. Hadoop was going to change the world. It kind of didn't, but we get the idea that it did. Kind of did, didn't. But it was good. It changed our world. It brought open source and the notion of low cost hardware into the big data game. And then the big data became so much more powerful around these new tools. But then the cloud comes in full throttle. It's like, wow, I can get horsepower. Got compute, can stand up, infrastructure analytics, all this data goodness starts to change. Machine learning then becomes the real utility that's showing this demand for using data right now. Not the setup, using data. This is a fundamental big trend. So I want to get you guys reaction. Do you see this evolving more cloud like? How do you guys see the trend in this as data science certainly becoming more mainstream, productivity, users to hardcore users. And then you got cloud native developers doing things like Kubernetes. We've heard Kubernetes here. It's like cloud is a data science. What's going on? What's your view of the market? So I came from a company that was in ad tech and we were built on big data. And in looking at how big data has evolved and the movement towards analytics and machine learning really being enabled by big data, people have rushed to build these solutions and they've done a great job. But it was always about what's the solution to my problem? How do I leverage this data? And they've built out these platforms. And in our context, what we've seen is that enterprises get to a certain point where they say, OK, I've got all these different stacks that have been built, these apps that have been built to solve my BI and analytics problems. But what do I do about how do I manage all these? And that's what I encountered in my last company where we built everything ourselves and then so wait a minute. But what we see in an enterprise level is fascinating because when I go to a large company, I go, we work with no SQL databases in Hadoop. And how much couch base do you have? How much Mongo, et cetera? The inevitable answer is yes and five of each. And they're getting to this point where I've got all this distributed data distributed across my organization. How am I going to actually manage it and make sure that that data is protected, that I can migrate to the cloud, or in a hybrid cloud environment? And all these questions start to come up at an enterprise level. We actually have had some very high level discussions at a large financial institution here in New York where they literally brought 26 people to the meeting. The initial meeting, this was literally a second call where we were presenting our capability because they're now at the point where it's like, this is mission critical data. This is not just some cool stuff somebody built off in one of our divisions. It matters to the whole enterprise. How do we make sure that data is protected, backed up? How do we move data around? And that's really the trend that we're tapping into and that the founders of our company saw many years ago and said, I need to, we need to build a solution around this. It's interesting, you think about network data as a concept or data in general. It's kind of got the same concepts we've seen in networking and or cloud. A control plane of some sorts out there. And where networking kind of went wrong is the management plane was different than the control plane. So management and control are huge issues. I mean, you bring up this sprawl of data. These companies are data full. It's not like, hey, we might have data in the future. They got data now. They're like bursting with data. One, what's the control plane look like? What's the management plane look like? These are all, these are technical concepts. With that in mind, this is a big problem. What are companies doing right now? What are some of the steps that are taking now to get a handle on the management? The data management is not just your grandfather's data management solution anymore. It's different. It looks different. Your thoughts on this challenge of management. So they're approaching the problem now. And that's our sweet spot. But I don't think they have, in their minds yet, come to exactly how to solve it. It's, there's this realization about we need to do this at this point. And in fact, doing it right is something that our founders, when they built the company, said, look, this problem of data management across big data needs to be solved by a dataware platform built on big data. So let's use big data techniques to solve the problem. All right, so let's, before we get into some of the solutions that you guys are doing, take a minute to explain what you guys are doing for the company, the mission, the value proposition, status. What do you guys do? How are people going to consume your product? I mean, take us through a quick introduction. Simple elevator pitch. We are enterprise data management, focused specifically on Hadoop and NoSQL. So everyone's familiar with the traditional space of data management in the relational space, relational world, very large market, very mature market. What we're tapping into is what John was just saying, which is you've got this proliferation of a Dupen NoSQL and people are hitting the ceiling because they don't have the same level of operational tools that they need to be able to mainstream these deployments, whether it's data protection, whether it's orchestration, whether it's migration, whatever the case may be. So what we do, that's essentially our value prop, is we're enterprise data management for a Dupen NoSQL. We help organizations essentially drive that control plane really around three buckets, data protection. If it's business critical, I got to protect it, okay? Disaster recovery falls into protection bucket. Good old stuff everyone's familiar with, but not in the Dupen NoSQL space. Orchestration is the second big bucket for us, which is I'm moving to an agile development model. How do I do things like automated task death? How do I do things like GDR, compliance management? How do I do things like cloud migration? John touched on this one before. Really interesting trend that we're seeing is using what our customer is doing, they're trying to create a unified taxonomy. They're trying to create a unified data strategy, which is why 26 people end up in the meeting. But in lieu of that, there's this huge opportunity because what they need, they know that it's got to be protected and they have 12 different platforms. And they also want to be able to do things like, well, I'm Mongo today, but I'll be Cosmos tomorrow. I'm Hadoop today, but I might be HD Insight tomorrow. I want to just move from one to the other. I want to be able to do intelligence. So essentially the problem that we solve is we give them that agility and we give them that protection as they're sort of figuring this all out. So I get this right. You basically come in and say, look, you can have whatever platform you want for your data, whether it's Hadoop and within those sequels, you have unstructured and structured data together, which makes sense. But protection specifically doesn't have to morph and get swapped out based upon a decision that they make. Well now we're focused specifically on Hadoop and NoSQL. So we would not be playing like, we're not the 21st vendor to be helping SAP and Oracle customers back up their data. It's basically, if you're Hadoop or NoSQL, that's the platform regardless of what Hadoop distribution you're doing or what. NoSQL distribution. So you don't have to change out your piece, what they do as they evolve and aren't. Correct, you're filling white space, right? Because when this whole movement started, it was like you were saying commodity hardware, and you had this idea of pushing code to data and oh hey, life is so easy and all of a sudden there's no governance, there's no data protection. Exactly, none of the enterprise. There's no business continuity, there's all this enterprise stuff. And then you heard for a long time, people were going to bring enterprise grade to Hadoop, but they really didn't focus on the data protection space. Correct, or the orchestration, either one of the big sections of those buckets, and you touched on just the last piece of that puzzle value-wise is on the machine learning piece. We do protection, we do orchestration, and we're bringing machine learning to bear to automate protection. Well the automate thing we hear a lot and that's a huge concern because the HDFS clusters need to talk to each other, right? There's a lot of nuances in Hadoop that are great, but also can create headache from a user human standpoint. Errors can get folded, I gotta write scripts. It creates a huge problem on multiple fronts. The whole notion of being eventually, being clustered in the first place, being eventually consistent in the second place, it creates a huge opportunity for us because this notion of being a legacy, we ask the question a lot, well you know there are a lot of traditional vendors who are just getting into this space and then what do you do then? That's actually good because it rises all boats, if you will, because we think we've got a pretty significant technology mode around our ability to provide protection and orchestration for eventually consistent clustered environments which is radically different than the traditional ones. I love the story about the 26 people showing the meeting, take me through what happened because it's kind of like one of the fish bowl. What do they do? Are they sitting there auditing? Are they taking notes? Are they really raising their hand? Are they tempering you with questions? What happened in that meeting? Tell us. So it's an interesting microcosm. What's happening in these organizations because as the various divisions and kind of like the federated IT structure started building their own stuff and I think the cloud enabled that. It's like basically giving the middle finger to central IT and say I can do all this stuff myself. And then the organization gets to this realization of like no, we need a central way to approach data management. So in this meeting basically, so we had an initial meeting with a couple of senior people and said we are going to talk about consolidating how we manage all this data across all these platforms. We want you to come in and present. So when we presented, there was a lot of engagement, a lot of questions. You could also see people still though there's an element of I want to protect my world and so this organizational dynamic plays out. But you know when you're a Fortune 50 company and data is everything, there's the central control starts to assert itself again. And that's what we saw in this. Because the consequences of not addressing it is what? Is potentially massive data loss, loss of millions, hundreds of millions of dollars. Data is the gold now, right, is the new oil. So the central organizations are starting to assert that. So we see that playing out and that's why all these people were in this meeting, which is good in a way because then we're not like, okay, we got to sell 10 different groups or 10 different organizations. It's actually being, so there's kind of this pullback to the center. What's happened in the NoSQL world of your perspectives on this? I mean, early on you had guys like Mongo took off because they were so simple to use and capture unstructured data. Now you're hearing everybody's talking about asset compliance and enterprise, great capabilities. That's got to be a tailwind for you guys because you're bringing in the data protection and orchestration component. But what do you see in that world? What do you guys support today and maybe give us a glimpse of the future? Sure, let me go. So what we see is, well, a couple different things. We are agnostic to the databases in the sense that we are definitely at Switzerland. We support all commerce, so it's follow the market share, if you will, Cassandra, Mongo, Couch, DataStacks, right on down the line on the NoSQL side. And what's interesting, they have all varying degrees of maturity in terms of what their enterprise capabilities are. Some of them offer sort of rudimentary backup type stuff. Some fancy, they have more backup versus others. But at the end of the day, their core differentiation, it's fascinating. They each have sort of a unique value prop in terms of what they're good at. So it's a very fragmented market. So that's a challenge, that's an opportunity for us, but it's a challenge from a marketplace standpoint because they've got to carve out their, they all want the biggest slice of the pie, but it's very fragmented because each of them is good at doing something slightly different. Yeah, okay. The situation described before is they've got, yeah, so they've got one of everything. So they've got 19 different backup and recovery processes approach, or they have nothing. Or scripting, which is massive now. So if they do have it, they've got a zillion steps associated with that and they're all scripted. And so their probability of a failure goes through the roof. Well, a human error, that's a fact. Well, a human error too is another problem, right? Absolutely. And you use the word tailwind and I think that's very appropriate because with most of these vendors, they've got their hands full just moving their database features forward, right? You know, we're the engagement fund. So when we can come in and actually help them with a customer who's now like, okay, great, thank you, database platform. What do you do for backup? Well, we have a rudimentary thing, we ship along with it, but there is one of our partners, Emanus, who can provide these, like robust enterprise, it really helps them. So with some of those vendors, we're actually getting a lot of partner traction because they see it's like, that's not what their strength is and they got to focus on moving their database forward. So I'll give you some stats. I'm writing a piece right now. Okay. Love stats. On the traditional enterprise backup recovery, but I wonder if you could comment on how it applies to your world. So these are research that David Floyer did and some survey work that we've done. On average, a global 2,000 organizations will have 50 to 80 steps associated with its backup and recovery processes. And they're generally automated with scripts, which of course are fragile, right? And they're prone to error. And it's basically because of all this complexity, there's a one in four chance of encountering an error on recovery, which obviously is going to lead to longer outages. And if you look at, I mean, the average cost of downtime for a typical global 2,000 companies between 75,000 and $215,000 an hour. Right. Now, I don't know, is your world, because it's data, it's all digital. It's worse. It's built, as I was going to say. It's worse. Is it probably higher end of the spectrum? All those numbers go up. All those numbers go up and here's why. All those metrics tied back to a monolithic architecture. The world is now microservices based apps and you're running these applications in clustered distributed architectures. Drop a node, which is common. I mean, think of, you know, you're talking about, you're talking about commodity hardware, commodity infrastructure. It's completely normal to drop nodes. It drops off, you just add one back in, everything keeps going on. If your script expects five nodes and now there's four, everything goes sideways. So the probability, I don't have the same stats back, but it's worse because the likelihood of error based upon configuration changes. Something as simple as that. And you said microservices was interesting too, is that not only is it just a data lake kind of idea of storing data in a dupe cluster. With microservices, now you're having data that's an input to another app. Check. So now, the level of outage severities, multiple, because there could be a revenue generating app. It could be some sort of recommendation engine for e-commerce or something. Something that's important. Like the data is right there because they're facing. Sorry, you can't get your bank balance right now. We can't do any transfers because the dupe cluster is down. I mean, this is pretty big. Yes. So it's a little bit different than saying, oh well, the cost to have a guy go out there and add a new server, maybe a little bit different. And this is the type of, those are the types of stats that organizations that we're talking to now are caring a lot more about. It speaks to the market maturity, if you will. Do you run into the problem of, you know, it's insurance, and so they don't want to pay for insurance, but a big theme in the traditional enterprises, how do we get more out of this data, whether it's helping manage, you know, I guess that's where your orchestration comes in. Cloud management, maybe cloud migration, maybe talk about some of the non-insurance value-add components and how that's resonating with customers. Yeah, so I'll jump in, but the non-protection stuff, the orchestration bucket, we're actually seeing, it comes back to the problem statement we just said before, which is they don't have, it's not a monolithic stack, it's a microservices-based stack. They've got multiple data sources, they've got multiple data types. It's sort of a, it's the byproduct of essentially putting power into division's hands to drive these different data strategies. So, you know, the whole cloud, let me double-click on cloud migrations is a huge value prop that we have. We talk about this notion of being data aware. So the ability to, I'm here today, but I want to be somewhere else tomorrow, is a very strong operational argument that we hear from customers, that we also hear from the SI community, because they hear it from the other community. And the other piece of that puzzle is you also hear that from the cloud folks, because you've got multiple data platforms that you're dealing with, that you need agility to move around. And the second piece is you've got the cloud, obviously there's a massive migration to the cloud, particularly with the Dupenose SQL workloads. So how do I streamline that process? How do I provide the agility to be able to go from point A to point B, just from a migration standpoint? So that's a very, very important use case for us. Has a lot of strategic value. Like it's coming, it's sort of the market's talking to us, like, this is simple, we have to be able to do this. And then simple things like, not simple, but automated test dev is a big deal for us. Everybody's moved to agile development. So they want to spin up, I don't want to basically, I want 10% of my dataset, I want to mask out my PII data, I want to spin it up on Azure. And I want to do that automatically every hour, because I'm going to run six builds today. Cloud certainly accelerates your opportunity big time. It forces everything to the table. Because you can't hide anymore. What are you going to do? You got to answer the questions. These are the questions. Okay, my final question I want to get on the table is, before we end the segment is, the product strategy, how are you guys looking at it as a SaaS, is it going to be software on-premise cloud? How does that look? How are people going to consume the offering? And two, opportunities, because you guys are a young growing company, you're kind of good timing. You don't have the dog more of the bag. Hadoop has changed a lot. Certainly there's a use case that everyone's getting behind, but cloud's now a factor. That product strategy. And then when you're in deals, why are you being called in? Why would someone want to call you? What are signs that would say, call you guys up? When would a customer see signals? And what signals would that be to give you guys a ring or a digital connection? You want to do that reverse order? I'll start with the customer, and then you can talk on the product. So the primary use cases are talked about back on recovery. There's also data migration and the test step. We have a big account right now that we're in final negotiations with where their primary use case is they're in healthcare and it's all about privacy, and they need to securely mask and subset the data. To your specific question around how are we getting called in, basically you've got two things. You've got the administrators, either the database architect or the IT or infrastructure people who are saying, okay, I need a backup solution. I'm at a point now where I really need to protect my data as one. And then there's this other track, which is these higher level strategic discussions where we're called in like the 26 person meeting. It's like, okay, we need an enterprise wide data strategy. So we're kind of attacking it both at the use case and at the higher level strategic. And obviously the more we can drive that strategic discussion and get more of people wanting to talk to us about that, that's going to be better for our business. And the stakeholders in that strategic discussion are whom obviously IT's involved, CIO maybe is their chief data officer and into healthcare. Yeah, database, enterprise architecture, head of enterprise architecture, various flavors, but you basically, it kind of always comes down to like two poles. There's somebody who's kind of owns infrastructure and then there's somebody who kind of owns the data. So it could be chief data officer, data architect or whatever, depending on the scale of the organization. They're calling you because they're full. They had to move the production workloads or they have production workloads that are kind of uncared for, un-nurtured. Is that the main reason they're in pain or are you the ass print? Are you more? No, there's like, we had a day loss and we didn't have any point in time recovery and that's what you guys provide. So we don't want to go through this again. So that's a huge impetus for us. It is all about to your point. It is mature. It's production workloads. I mean, the simple qualifying, are you running a doop or no sequel? Yes. Are you running in production? Yes. You have a backup strategy. Sort of tip of the spear. Now to just briefly to answer your question before we run out of time. So it's not a SaaS based, we're a software defined solution. We'll run in bare metal. We'll run in VMs. We'll run in the cloud, Azure, Google, whatever you want to run on. So we run anywhere you want. We're software defined. We'll use any storage that you want and basically it's an annual subscription base. So it's not a SaaS consumption model. That may come down the road but it's basically an annual license that you buy, deploy it wherever you want. Customers choose what they do. Basically customers can do, it's completely flexible. But back to your answer. I want to go back to something you said. You said they didn't have a point in time recovery. What their point in time recovery was their last full backup or they just didn't have one? Or they just didn't have one? All of the above. We've seen both. They said there's a market maturity issue as something to look at. Yeah, because it's like, oh it's replicated. We see a lot of people. We hear that a lot. It's clustered. You know, I just replicate my data. And replication is not our. And truth be told, my old company, that was our approach. We had a script but still it was like, and the key thing is, even if you write that script, as you pointed out before, the whole recovery thing. So, you know, having a recovery sandbox is really important. They didn't think about this when they designed everything. Exactly. Because it was an afterthought. It was all a rush to extract the value from the data. And show the use case. Prove it out. A dupe's real. We stood it up. And history is repeating itself in that regard. If you go back to the relational space, there's a very interesting correlation to the delta between the database platforms and the data management stuff. It's logical. Hence the architects are involved. They're coming in to, okay, let's look at this in a big picture. Exactly right. What's the recovery strategy? How are we going to scale this? Exactly. It's just a product question. So your granularity for a point in time is you offer a bunch of stuff? Any point in time. Any point in time is varying. We'll have more news on that in the next couple of weeks. I'm at this data here inside theCUBE, hot new startup growing companies, really solving a real need in the marketplace. You're kind of an aspirant today, but growth opportunity as they scale up. So congratulations. Good luck with the opportunity. It's theCUBE bringing you live coverage. Here's part of CUBE NYC, our ninth year covering the big data ecosystem starting originally 2010 with the Duke World. Now it's AI and machine learning. Hadoub cluster is going into production. Guys, thanks for coming on. I really appreciate it. This is theCUBE. Thanks for watching day one. We'll be here all day tomorrow. Stay with us for more tomorrow. Be right back tomorrow. I'll see you tomorrow.