 Live from Las Vegas, it's theCUBE. Covering Edge 2016, brought to you by IBM. Here's your host, Dave Vellante. Welcome back to IBM Edge, everybody. This is theCUBE's fifth year covering IBM Edge. We're at the inaugural Edge five years ago in Orlando. Mark Lindster is here, and he's joined by Lenly Henserling. Mark is the Senior Vice President of Product Development and Lenly is the Senior Vice President of Product Management and Strategy at EDB Enterprise Database. Gentlemen, welcome to theCUBE. Thanks for coming on. Thank you. Okay, who wants to start? Enterprise Database. Tell us about the company and what you guys are all about. Well, the company's been around for a little over 10 years now, and our job is really to give companies the ability to use Postgres as the platform for the digital business. So think about this. Postgres is a great open source database, great capabilities for transactional management of data, but also multi-model data management. So think about standard SQL data, but think also about document oriented, think about key value pair, think about GIS. So a great capability that is very, very robust has been around for quite a few years and is really ready to allow companies to build on them for the new digital business, but also to migrate off their existing commercial database that are too expensive. So what's the history of Postgres? Can you just sort of educate me on that? So the same roots back with system R where DB2 came from, Oracle came from, so Berkeley, that's where the whole thing started out. Postgres is really the successor to Ingress, right? And then it could turn into Postgres QL and has been licensed under open source license, the Postgres QL license since 1996. And it's a very, very vibrant open source community that's been driving forward for many years now. And in our view is the best available relational and multi-model database today. What's the main spring of relational database management systems essentially is what we're saying. So and Lenly, from a product standpoint, how do you productize that? You know, you got open source. Yeah, so open source really, you know, companies that have a distribution of open source for a database and operating system or whatever that the open source company most people are acquainted with is Red Hat and Linux, right? And so we do the same thing that they do but for Postgres database. We take the distribution, we add testing, we add some other functionality around it so that you can run Postgres responsibly as Mark likes to say. So high availability capability, failover management, replication, a backup solution. And instead of leaving it as an exercise for a customer who wants to use open source, we test all this together and then we validate it and we give them a complete package with documentation and services that they can access to help them, you know, be successful with it. So if Michael Stonebreaker were sitting right here and say, Michael, what do you think about Postgres? He say, I had to start Vertica because we needed a new way. Yet, sort of Postgres SQL is sort of the killer, it remains the killer platform in the industry, doesn't it? Why is that? I mean, it's interesting when you talk to guys like Stonebreaker, it's sort of dogma almost but yet customers, you know, talk with their wallet. Well, and it is, I mean, he did a very, very nice job architecting it, right? It is a database that is extensible. The reason we had the first JSONB or document-oriented implementation in the relational database space is because it was designed to make it easy to add new capabilities, new data types, new indexes, et cetera, into the same transactional model. That's why we have JSONB, that's why we have PostGIS, that's why we have KeyValuePair. So it was really well architected and when you think about who else, not just Vertica has taken this engine, you know, it is in the TZ, it is in a bunch of other platforms. Green Plum, yes. So it's a really robust architecture, very, very nicely designed and it just does the job and it does it really well, which is what you want a database to do, right? It's not that exciting, but it's really stable, it really works, the data is still there tomorrow. That's really what the requirements are. And to translate a little bit, Mark mentioned PostGIS, which is geospatial capability for the Postgres database. And so we distribute that along with Postgres and test it so that it works. And he mentioned HStore. So that's how you can actually store Internet of Things data really well into Postgres. And when we talk about SQL, no SQL databases, so they're document databases. And the ability to have personalizations at the same level you can in a documented or in a database, but in a structured SQL database, are the kinds of things that have been added to Postgres over the years. And again, it's the cause of the basic architecture that Stonebreaker put in place is an object relational database. It's so interesting to look at the history of database. And as I talk about Stonebreaker, he's been on a number of times, he's just fascinating to listen to one of the fathers of this industry. But 10 years ago, database was like such a boring topic. And now it's exploded, right? I mean, and now you got Amazon, going after Oracle, Oracle fighting the good fight, so many SQL databases coming in, SQL becoming the killer big data app, if you will. Why all of a sudden database gets so interesting? Yeah, so what happened was application models changed. Led by Facebook, led by Amazon and Google, right? They said, let's refactor the applications and let's refactor the way we handle storage. Okay, and that led to the rise of a polyglot of databases. This is what a lot of people are saying. So you have fit for purpose solutions, and you may have three or four or five of them in your overall architecture. One thing about Postgres is we're able to, because of the data type support that Mark mentioned, fit into that well. We don't try and do everything. So if somebody says, I'm going to use Mongo for data capture, or I'm going to use Cassandra for capturing my internet of things data. We have what we call foreign data wrappers in the Postgres world. We call them just enterprise DB adapters, but two Mongo, two Cassandra, two Hadoop, and can do bi-directional data there and just keep that data at rest over there in the other world, but be able to project a relational scheme on to it. We can push our data into those. We've got a great use case we've been talking about with a customer who had over a petabyte of data. And in the past, what you do is you'd go buy an expensive archiving solution and add that to it. Now you just use Hadoop distributed file system, push the data off there as it ages and have a foreign data wrapper that allows you to still query that data when it's out of your basic operational data set and move forward. Can I call that a connector or is that? A connector, yeah, that's not a good idea. And it's interesting because, I don't know if you guys remember Hadapt probably, right, they came out there with the connector killer and it failed. Seems like connectors are just fine. Yeah. And one of the really interesting things is we call it data federation, right? With the philosophy here is, leave the data where it is. I mean, there is some data that should live in Hadoop or Cassandra, right? If I'm doing an e-commerce site with transactions and click streams, well, the click streams really should live in Hadoop. That's the natural place for them. The transactions should be in a transactional database. With a foreign data wrapper, I can run queries without moving the data that will allow me to say, well, before you bought the brown teddy bear, which pages did you look at? And I can do that integrated system and I can do a fit for purpose architecture and that's what we think is really exciting. And that's fundamental to this new sort of programming or application model that you were talking about, is move five megabytes of code to a petabyte of data as opposed to moving data, which we know has gravity and speed of light issues and so forth. Well, thank you for that little brief education. Appreciate it. So let's get into your business now, your relationship with IBM, what customers are doing. You mentioned IOT data, so talk more about your business and your relationship with IBM and what you guys are doing for customers. There are a couple of things. We mentioned Oracle, right? And there's all the new databases and then there's your, I don't know, dare we say legacy proprietary databases as well. And people are looking to become more efficient in how they spend. So we've done another thing with Postgres. We've added Oracle compatibility in terms of data types. So we support all the data types that Oracle does and we support PLC. They're sort of variant of stored procedure language and have implemented a lot of the packages that they have as well. So we can migrate workloads from Oracle over into an open source based solution and give a lot of cost effectiveness options to customers. So this is a way that I can sort of have Oracle license, database license and maintenance avoidance, where possible. Yeah. And where it makes sense, where it makes sense. I'm going to keep, but let's face it, the number one cost component of a TCO analysis of an Oracle customer is the database license and maintenance cost. It's not the people. One of the few examples I can think of where that's the case. It's always the people cost. IT is very labor intensive, but for an Oracle customer, it's the database license, because they're licensed by core. Cores are going through the roof. It's great for Oracle's business. Although, I mean, when you agree that Oracle, I think sees the writing on the wall but the SaaS is really the sort of new control point for the industry. And you see the acquisition and net suite and competition with Workday and the like, but the database remains the sort of heart of their business. And really, it's movement to the cloud, both private cloud and public cloud. And so we've been doing work there. We've had a public cloud database as a service solution on Amazon for what? Four years. Four years, Mark. And have gained a lot of experience with that. And we were running that, sort of running a retail, you can license the database and we'll provision it there. And so what we've done recently is changed our perspective and said, let's put this into the hands of customers and let them stand up their own databases as a service, but also do it in a way that they can choose what workloads should go to Amazon and what workloads might go to their private cloud built on OpenStack and be able to arbitrage that, if you will, because they now have a way to provision the databases and make a choice about where to put it. So that's a bring-your-own-license model that you just talked about? A bring-your-own-license model, or? Or you're in the marketplace? We're in the marketplace on Amazon, where we can supply it that way, but customers have shown a preference for bring-your-own-license. They want to make the best enterprise deal they can with a vendor like us or whoever else, and then have control over it. I mean, Amazon obviously wants you to be in the marketplace, but I won't even mention it, but I talked to some CEOs of database companies and they say, you know, we're in the marketplace, but when we get on the marketplace, next thing you know, Amazon's pushing them toward DynamoDB, and now Amazon's come out with Aurora and Oracle Migration and the intent to go after that business, so Amazon's moving up the stack and you've got to be careful. They are, but the thing about Amazon is that they're a pure play in the cloud company, okay? And all of the data shows that it's like a mix. It's going to be a hybrid cloud. Half the companies, 80% of the people in the cloud are going to be on-prem, still continuing their journey through virtualization, let alone go into the cloud. But you know, we want to be something that lets them put what they want in the public cloud and lets them manage on the private cloud in the same manner so that they can provision databases with a few clicks, just like they do on Amazon, but do it in their data center. And you're doing that with SoftLayer as well, or not yet? Not yet, not yet. We've built this provisioning capability ourselves and it came out of the work we did putting up databases on Amazon. So what are you guys doing here at Edge? Edge's kind of infrastructure show. I mean, database is infrastructure. We're talking about our work with Power. Power is a big partner for us. I mean, Power is, I think, very, very interesting for database customers because of the much higher clock speeds and the capabilities that the power processor has, right? When I'm looking at Power, I get more oomph out of a single core, which really, for a database customer, is very, very interesting, right? Because all databases are licensed by core. So it's a much better deal for the customer. And specifically for Postgres, Postgres scales very well with higher clock speeds. So by having, let's say, by growing performance, not by adding more cores, but by making this individual course faster, that plays very, very well to the Postgres capabilities. Okay, so you are a power partner, one of the part of that ecosystem that IBM's appealing to to grow the open power base. And what kind of workloads are you seeing your customer's demand? And where are you having success? Across the board, I mean, database is mostly, infrastructure capability, so there's a lot of interest that we're seeing that, for all kinds of applications, really. What's the typical power customer look like these days? You got some Oracle, you got some DB2, you guys are running on there. What's the mix, kind of, in a picture for us? Yeah, I think that the typical power customer is the typical enterprise company, you know, and what they're, it's a little bit of everything, but you know, one of the key things is that people are also looking at what they've got and the skills they have in place. You were talking about people costs, right? Yeah. And their understanding of management, their understanding of how to manage the relationship with the vendor even, and then saying, look, how can I move into the new world of digital transformation and start, you know, my own private cloud options and things like that in an efficient way that makes efficient use of hardware I have in place and has a growth curve in new hardware that's coming out that fits my workloads and the profiles that Mark was talking about. Yeah, and also the resources, right, which is very interesting when we look at these new digital applications with Postgres, because you can do so much in Postgres from geographic information systems to document-oriented to key value, right? But you can do that with your existing developers through existing DBAs. They don't need to go to school to learn a new database, right? And that's also a very, very interesting capability. So you can use your existing team to do new stuff. Yep. What's happening in IoT? What problems are you solving there? Where's the data? Sensor data collection, right? Really interesting because sensor data tends to come in all different forms, right? I mean, we have a customer who collects temperature data, but the sensors are all, you know, sending different data packets. So because we can do document-oriented or key value, we can easily accommodate that, right? In the old days with the relational model, I had to do all kinds of tricks to sort of stuff all that into a relational table. My table would be almost empty at the end because I'd have to add columns for every vendor, et cetera. Here, now I can just put all that into the same format and, you know, provide it for analysis. So that's a really interesting capability. And it's interesting too because, you know, we've got really strong geospatial data support and the intersection of that with IoT is a big deal. I mean, you know, they track your iPhone, they know where we are, they know what's going on. That's sensor data. They know which lights and which building, which, you know, louvers that are controlling HVAC are malfunctioning or not. They want to know specifically where it is, not just what the sensor is. And some of that stuff moves around, right? And it gets replaced in a new place in the building and such. So we're well set up to handle those types of workloads. It was interesting when IBM bought the weather company and they, you know, you thought, okay, great, they're getting all these data scientists and weather data, that's cool, they can monetize that, but it's an IoT play, isn't it? Because they talk about, you know, sensors. It's reference data, it's reference data for other company specific IoT plays to have a broader set of sensors out there in their region and understand what's happening with weather and things and then play that against what their experience is and managing their buildings, their manufacturing processes, everything. So what's the engagement model? Customer, I want to do business with you. How do I do it? How do I engage? Well, I mean, a lot of our business is direct with us, right? To others, through partners and a lot of customers come to us because they want to get off legacy systems. But really what they do is once they understand the database and the capabilities, they say, okay, yeah, you can do the Oracle stuff, right? But what I'm really going to do with you is my new things because that's really exciting and it helps me kind of put a lid on the commercial license growth, right? So yeah, maybe I'm not going to get off it, but I will stop growing it. So I will start doing my new stuff on Postgres. Whenever I modernize something, Postgres is going to be my database of choice if I already open up an application with its whole stack, right? This is one of the changes I'm going to make and then the database as a service is very, very interesting. So these four entry vectors and then what happens is quite a few customers after a short time, when they started with projects or applications, they end up making Postgres one of their database standards. Not the only one, right? But they make it one of the database standards so it gets into the catalog and every new project then has to consider Postgres. It's interesting. There's a space created as Microsoft sort of put all their wood behind the air of becoming a competitor to high-end Oracle and with this last release, they probably got in there, arguably. But they've also raised their prices too, right? And they've made the solution more complex and so there's this space that was vacated for like a ton of workloads and Postgres fits in there just about perfectly and we see enterprise after enterprise come to us with a sheet that says, you know we're going to get some of this no-sequel stuff. We're going to keep Oracle or DB2 over here for these really high-end things. You know, run my financials, run my sales order processing, my manufacturing. And then we got this space in here. We got a slot for a relational database and we want to go open source because of the cost savings, because of other factors, you know. It's ability to grow and not be bound to, hey, what if the vendor decides they're going to go for a newer cooler thing and make me upgrade and I want to stay there and know that there's still being an investment made and so there's a vibrant community around it and it just fits that slot perfectly. You got to pay for that digital transformation and all these IOT initiatives, you can't just keep pouring it down to database licenses, so. All right, gentlemen, we have to leave it there. Thanks very much for coming to theCUBE and appreciate the time. You're welcome. All right, keep right there, everybody. We'll be back with our next guest. This is theCUBE, we're live from IBM Edge 2016. Right back.