 Live from the Moscone Convention Center in San Francisco, California, it's the Q at Oracle Open World 2014. Brought to you by headline sponsor Cisco Systems with support from NetApp. And now here are your hosts, Stu Miniman and Jeff Frick. Hi, welcome back. I'm Jeff Frick. You're watching the Q. We are live at Oracle Open World 2014. Day three of wall-to-wall coverage. We've been here extracting the signal from the noise. There's a lot of buzz, a lot of energy. Larry has basically said cloud is good, so that's official now. So we're excited about that. I'm joined in this next segment by my co-host. I'm Jeff Kelly with wikibon.org. So yeah, as you mentioned, Jeff, cloud, of course, is the biggest buzzword here at the show. But I'd say probably the second biggest buzzword is big data. So in this segment, we're going to talk with Dave Segler, who's the director of product management for NoSQL database, Berkeley DB database and mobile server for Oracle. Dave, welcome. Thank you. That's a big title. So tell us a little bit. It's your first time on the Q. Why don't you tell our audience a little bit about kind of your role in Oracle and then we'll kind of get into Oracle's approach to NoSQL. Sure. I'm the director of product management for all of those products. And essentially, I came on board with Oracle after the acquisition of Sleepy Cat and the acquisition of the Berkeley DB technology, which is the basis for our Oracle NoSQL database product. So I've been doing key value pairs probably for about a dozen years now, starting with Berkeley DB and moving up to Oracle NoSQL database. As part of our organization, we also inherited the Oracle database mobile server, which has the ability to link mobility and big data via synchronization engine. Okay, that's a really interesting topic. But why don't we start with NoSQL and Oracle's approach? Because obviously, Oracle is known as the database company. Right. And I should say the relational database companies. Now you have NoSQL coming into the picture. We've seen it over the last several years. We were talking off camera a little bit about how this is, to us in the industry, it seems like it's been going on for a while, but it's still very early. But if you could just kind of articulate what is Oracle's thoughts approach philosophy towards NoSQL, you know, coming from that long history in relational technology. So NoSQL is definitely a component to big data and is part of the strategy of storing and managing very fast, but simple operations over simple data. And so from Oracle standpoint, they recognize the value of NoSQL in a big data solution. And that's why we started the project about five years ago, launched the product in 2011. And kind of the vision is NoSQL as it matures as a product and a technology will be very much part and parcel of what Oracle views as their big data strategy. And clearly part of that is the Oracle Big Data SQL, which has the ability to query not only relational data, but data in Hadoop and data in NoSQL as well. Another interesting point. So SQL on Hadoop has been a big topic in the industry. Talk about the value that Big Data SQL from Oracle brings that maybe you're not potentially seeing some of the other solutions, whether it's from the open source community, whether from competitors slash partners like Cloudera with Impala. What does, what makes Big Data SQL from Oracle different than some of the other approaches? Primarily decades of implementation of research and technology, the Oracle Big Data SQL leverages the information and the expertise we gained with Exadata and how to essentially make SQL queries extremely both efficient and distributed. So it's not just about running SQL across a Hadoop installation, for example. It's about creating and enabling an engine that can do proper SQL processing in a distributed fashion. So it helps break down those silos because it can query data across both SQL, Hadoop, and your relational technology. Exactly. From a centralized query processing standpoint. Okay. Now, does that, so does that require an investment in Big Data Appliance specifically? If you want to use Big Data SQL across these different systems? Does it work just across Big Data, the Big Data Appliance and Exadata? Can it query other systems that are maybe not Oracle or from other areas or? Today it works on the Big Data Appliance and Exadata. It's part of our value of pre-engineered systems and can access data in other repositories as well. Okay. So let's again, let's just take a little bit of a step back and talk about the fundamental approach. So I think we think about no SQL and at least what your competitors are trying to position it as is, you know, it doesn't require proprietary hardware. You can scale this out and come out into machines. Now Oracle, obviously known for very powerful but sometimes expensive machines. How does that, how does Oracle get into this market in a way that makes sense financially for Oracle, but also for customers who one of the value propositions of no SQL is, it's essentially less expensive and doesn't require as big of an upfront investment. So the legacy of the product and my legacy personally in, in the industry is, is open source. WorklyDB was an open source product is continues to be an open source product. And so Oracle no SQL database runs on both the Big Data Appliance pre-engineered systems and optimized for that environment, but also runs on commodity software and Oracle no SQL database comes in two flavors. The community edition is open source under the HEPL license. So it's available for free. And the enterprise edition comes with an Oracle license. So I think what Oracle looks at is we need to enable storage and manipulation of large amounts of data in a no SQL database. And we need to leverage how that how we monetize that by offering, for example, efficient implementations on pre-packaged hardware with integration into the rest of the Oracle universe. Certainly for Oracle customers, one of the important value ads is that we care about integration. No SQL no SQL is not a silo of data. No SQL is integrated with the rest of the application infrastructure. So unlike some of our competitors, we are interested in how does this integrate with Oracle coherence? How does it integrate with event processing? How does it integrate with the Oracle database? How does it integrate with enterprise manager? And for Oracle customers, I think that's part of the value ad that we bring to the table. So David, what if I could follow up? There's a lot of talk about Oracle engineered systems, right? And Oracle software runs best on Oracle hardware, but I wonder if you can give us some hints and some secret sauce and you know, why, you know, what are some examples of things you could do in an engineered system that you couldn't if you just say running on on standard commodity hardware standard commodity. One of the one of the main things is that we can make the communications interface, for example, very efficient. So in Finland is a huge benefit that you don't normally find in commercially available or commodity systems. So one of the key aspects, especially of no SQL and big data is you're constantly exchanging data between the different nodes in the cluster. If you're building that cluster based on tightly integrated hardware with high throughput intercommunications, you can do things both in parallel and at a speed that's just not achievable on your standard vanilla commodity hardware platform. It's a good example. So let's talk about the relationship with Cisco. So you've got the kind of the red stack and the red box, but you've also got a relationship with Cisco. Yeah, talk about what you guys are doing there around no SQL and UCS offer. So we've had a long standing relationship with Cisco. In fact, Cisco was one of our earliest partners in benchmarking and testing Oracle NoSQL database on platforms that even predated the big data appliance. In fact, we have run the largest cluster tests that we've done on Cisco UCS platform. So we have a long tradition of working directly with Cisco to say, okay, what are what is it the Cisco customers need and are looking for from UCS? And how does UCS make installation and operation of Oracle NoSQL more efficient? In particular, Cisco is an expert in interprocess communication and communication between nodes. So one of the values that Cisco brings to the relationship is understanding what's the bandwidth requirements, for example, of communication between nodes, especially important in clustered environments. So maybe put a little more color on that. So maybe if you could identify a couple of use cases where that's critical, and maybe some of the customers that you think this partnership is going to be most attracted to. So in a no SQL system, basically, you have two ways that you're exchanging data every every second inside the cluster. So on the one hand, you're communicating from the client to the server. So there's a lot of data flowing there because we're managing lots of data and lots of queries. But also any updates are getting replicated across the cluster. So you've got communication going on between the servers under the covers. So unlike a traditional relational database where most of the traffic is client server in the no SQL environment, you've both got client server and server to server communication, as well as things like topology management gossip protocols, administrative commands that require not only high bandwidth, but balanced bandwidth between the nodes. So let's get into some of the interesting use cases. What are you seeing out there with some of Oracle's early no SQL customers? What are some of the things that they're doing? Have you seen any patterns in terms of the type of customer? Obviously as we talked about, it's still a still early day. So I'm guessing it's going to be fairly, you know, it's early adopters. It's more advanced or more risk taking organizations that are willing to experiment with new technology. But what are you seeing out there in terms of some of the killer applications? So one of the first things I've seen is the first couple of years we were working in the no SQL space, I spent a lot of time educating people about what no SQL is and what the technology can do. The last couple of years have been more about customers coming to us and saying, we have a no SQL problem. How does your product address that problem? So we've gone from experimentation to implementation, which as a technologist is always very interesting because we're starting to solve real world problems. A couple of the things that we're seeing in common amongst our customers is, first of all, nobody's using no SQL in isolation. Every single customer that we have is using a relational database, usually Oracle, Hadoop, and no SQL in their overall solution. And they're combining them in interesting ways in leveraging the capabilities and capacities of each technology to the best of their ability to solve their technical and application challenge. In terms of application use of the product, what we see is a lot of, for example, product recommendation websites. NTTDocomo just launched a site this year called DeMarket that allows all of their mobile cell phone users to get product recommendations for digital content. So when you log on to the DeMarket page, you get a series of product recommendations based on who you are, based on your past purchasing patterns, what promotions are available, what products are available in the product catalog. And it's this operationalization of market segmentation. How do I identify and provide value to my customers and increase my bottom line? How do I build a service that does both? And in both their case as well, several others that we've seen, they combine an analytics engine and analytics capability with no SQL, which gives them the platform to operationalize and serve that information in a personalized manner to very large sets of customers concurrently. And I think that that's also a good illustration of the integration requirements because I'm guessing that's pulling on data, the analytics is pulling data from multiple systems. Yes. So you've got customer data in your systems of record, you've got maybe social data coming in, who knows where it could be coming in. So the integration between the different data sources, being able to do the analytics, run the analytics, and then feed that into essentially a real-time or near real-time system. It's a big data integration challenge. It is. And it's a question that several customers have addressed in terms of, I have results. I have results from these analytics. Where do I put them? How do I operationalize those results? And what we've found in several customer use cases is they say we have the analytics capability, be it Oracle Advanced Analytics, be it some other product, be it something they wrote in R. But how do we make that available to our customers in the form of a service that either we charge for or generates more revenue? And that's where NoSQL comes into the picture. The sources, the analytics bring in different sources. They take the conclusions, put those in a platform that's both scalable and expandable and provide that out to their customers as a service. A great example. And I think, you know, what's interesting, what's happening in the market is you're seeing kind of, I think the old world was kind of, you know, it was OLAP and OLTP kind of separated, kind of siloed. And you could sort of think of the corollary as NoSQL and Hadoop. But we're starting to see those analytics and transactions start to merge a little bit. So you're building transactional or operational applications that are leveraging real-time analytics. And that's where we're seeing NoSQL in particular playing an important role. So things like Hadoop still really great for that large-scale historical analysis. There's some overlap with some of the real-time stuff people are trying to do in there with Spark and some other things. But the ability to kind of combine both transactional and analytic capabilities to build smart applications, whatever the technology underneath is, is really where it gets exciting. Well, I think the challenge with big data in general is how do you take that data and turn it into useful, leverageable information, useful to your customers, leverageable for your business? And how do you take that result of your analysis, that result of combining all those data sources and turn it into something that people can actually use? And so, as you said, it's that combination of bring in multiple sources, bring in multiple technologies, and then find the platform that will operationalize that for you. So it begs an interesting point, kind of more philosophical, where Jeff just said, it used to be everyone kind of siloed, we do this, you do that, and you know, and the reality is different, different, different pieces do different things well. So talk a little bit about, you came from the open source, you know, kind of open source inside of Oracle, in terms of, you know, the reaction to it, kind of the adoption of it, and it almost sounds like really an embracing within a solution set that leverages the best of what those things deliver. I wonder if you can just give some perspective. People don't think of Oracle and open source necessarily in the same breath. And it's a funny, it's a funny impression that people have because Oracle owns Java, which is open source. That's true. They acquired MerkleyDB back in 2006. MerkleyDB is still open source. They haven't changed the license. So Oracle contributes to Linux. Oracle has a large open source community and people dedicated to open source within the company. And I think Oracle as a sales organization was historically accustomed to saying, oh yeah, that open source stuff, we don't make money there. But I think what I've seen over the last several years is definitely a change in perspective of no, it's not about just the red stack. It's about solving customer problems which may involve components that are licensed and components that are open source. And Oracle is very much embracing that as part of a solution set. Which is interesting because I think it is that sales culture that drives people's perception of what Oracle is. And so for them to start to take that in print, to take that forward to the market should have a significant impact on the perception. And I think that happens for two reasons. Part of it is because there's a dedicated open source community within the development organization. But also based on what they hear from customers. So if your customer says, I'm interested in using an open source new technology like Spark or NoSQL or Hadoop. And the only answer you give them is, oh, here's more exadata. Well, that's not a conversation that's going to last very long. So the customers are helping educate us as well. We're reacting to the changes in the marketplace in a way that we recognize that customers are looking for solutions that are not just one stack, but picking the technology that best solves their application challenges. So we're close on time. But I just I would love to get your take on just the competition, the market. Obviously, we've got all these startups out there. Obviously, we want to hear, is Oracle going to make any acquisitions? What's the plan? But I'm guessing you're not going to answer that. But just how do you see this shaking out, considering your history, considering what we see in the database space, the relational database space? Do you see consolidation coming at some point not necessarily Oracle making acquisitions? But how do you see it shaking out over the longer term? So right now, if you go to noSQL database.org or any of the other sites and you start to do research on NoSQL, there are well over 100 vendors selling NoSQL type solutions. If you look at the market data, the market analysis is it's a market somewhere between 200 and $400 million. I'm sorry, that's not enough space for 100 companies. So over the next several years, it's clear that there is going to be consolidation in the space. There are lots of startup companies with lots of interesting technology, but they have to be funded, right? That there has to be a revenue plan that gets them return on investment for venture capitals. So I think the reality of the marketplace is that we've seen an explosion of innovation, an explosion of implementation techniques and what we're going to see over the next two, three, five years is those companies that have the technology that solves customer problems predictably and is manageable in a production environment are the companies that are going to continue forward and the companies that have interesting technology but not necessarily manageable in production or solving specific customer problems are the ones that are going to have to win away. I mean eventually you can't have 100 companies in a $400 million market space. That's true. Although we, you know, we as we've been sizing the market, we do expect it to grow significantly, but yeah, 100 companies is quite a bit. And you saw the database space kind of the consolidation and there's space for a handful of players, but perhaps not perhaps on 100 triple digits is a bit right. But anyway, I wish we had more time, but David, thank you so much for joining us on the cube. Love to have you back on again as we continue to cover this really important topic. Thanks everybody for watching the segment. We'll be right back with our next guest here live at Oracle Open World.