(bright upbeat music) >> Hello everyone, and welcome to this exclusive Cube Conversation. We have the pleasure today to welcome, Wim Coekaerts, senior vice president of software development at Oracle. Wim, it's good to see you. How you been, sir? >> Good, it's been a while since we last talked but I'm excited to be here, as always. >> It was during COVID though and so I hope to see you face to face soon. But so Wim, since the Barron's Article declared Oracle a Cloud giant, we've really been sort of paying attention and amping up our coverage of Oracle and asking a lot of questions like, is Oracle really a Cloud giant? And I'll say this, we've always stressed that Oracle invests in R&D and of course there's a lot of D in that equation. And over the past year, we've seen, of course the autonomous database is ramping up, especially notable on Exadata Cloud@Customer, we've covered that extensively. We covered the autonomous data warehouse announcement, the blockchain piece, which of course got me excited 'cause I get to talk about crypto with Juan. Roving Edge, which for everybody who might not be familiar with that, it's an edge cloud service, dedicated regions that you guys announced, which is a managed cloud region. And so it's clear, you guys are serious about cloud. These are all cloud first services using second gen OCI. So, Oracle's making some moves but the question is, what are customers doing? Are they buying this stuff? Are they leaning into these new deployment models for the databases? What can you tell us? >> You know, definitely. And I think, you know, the reason that we have so many different services is that not every customer is the same, right? One of the things that people don't necessarily realize, I guess, is in the early days of cloud lots of startups went there because they had no local infrastructure. It was easy for them to get started in something completely new. Our customers are mostly enterprise customers that have huge data centers in many cases, they have lots of real estate local. And when they think about cloud they're wondering how can we create an environment that doesn't cause us to have two ops teams and two ways of managing things. And so, they're trying to figure out exactly what it means to take their real estate and either move it wholesale to the cloud over a period of years, or they say, "Hey, some of these things need to be local maybe even for regulatory purposes." Or just because they want to keep some data locally within their own data centers but then they have to move other things remotely. And so, there's many different ways of solving the problem. And you can't just say, "Here's one cloud, this is where you go and that's it." So, we basically say, if you're on prem, we provide you with cloud services on-premises, like dedicated regions or Oracle Exadata Cloud@Customer and so forth so that you get the benefits of what we built for cloud and spend a lot of time on, but you can run them in your own data center or people say, "No, no, no. I want to get rid of my data centers, I do it remotely." Okay, then you do it in Oracle cloud directly. Or you have a hybrid model where you say, "Some stays local, some is remote." The nice thing is you get the exact same API, the exact same way of managing things, no matter how you deploy it. And that's a big differentiator. >> So, is it fair to say that you guys have, I think of it as a purpose built club, 'cause I talk to a lot of customers. I mean, take an insurance app like Claims, and customers tell me, "I'm not putting that into the public cloud." But you're making a case that it actually might make sense in your cloud because you can support those mission critical applications with the exact same experience, same API, same... I can get, you know, take Rack for instance, I can't get, you know, real application clusters in an Amazon cloud but presumably I can get them in your cloud. So, is it fair to say you have a purpose built cloud specifically for the most demanding applications? Is that a right way to look at it or not necessarily? >> Well, it's interesting. I think the thing to be careful of is, I guess, purpose built cloud might for some people mean, "Oh, you can only do things if it's Oracle centric." Right, and so I think that fundamentally, Oracle cloud provides a generic cloud. You can run anything you want, any application, any deployment model that you have. Whether you're an Oracle customer or not, we provide you with a full cloud service, right? However, given that we know and have known, obviously for a long time, how our products run best, when we designed OCI gen two, when we designed the networking stack, the storage layer and all that stuff, we made sure that it would be capable of running our more complex environments because our advantage is, Oracle customers have a place where they can run Oracle the best. Right, and so obviously the context of purpose-built fits that model, where yes, we've made some design choices that allow us to run Rack inside OCI and allow us to deploy Exadatas inside OCI which you cannot do in other clouds. So yes, it's purpose built in that sense but I would caution on the side of that it sometimes might imply that it's unique to Oracle products and I guess one way to look at it is if you can run Oracle, you can run everything else, right? Because it's such a complex suite of products that if you can run that then it it'll support any other (mumbling). >> Right. Right, it's like New York city. You make it there, you can make it anywhere. If I can run the most demanding mission critical applications, well, then I can run a web app for instance, okay. I got a question on tooling 'cause there's a lot of tooling, like sometimes it makes my eyes bleed when I look at all this stuff and doesn't... Square the circle for me, doesn't autonomous, an autonomous database like Autonomous Linux, for instance, doesn't it eliminate the need for all these management tools? >> You know, it does. It eliminates the need for the management at the lower level, right. So, with the autonomous Linux, what we offer and what we do is, we automatically patch the operating system for you and make sure it's secure from a security patching point of view. We eliminate the downtime, so when we do it then you don't have to restart applications. However, we don't know necessarily what the app is that is installed on top of it. You know, people can deploy their own applications, they can run third party applications, they can use it for development environments and so forth. So, there's sort of the core operating system layer and on the database side, you know, we take care of database patching and upgrades and storage management and all that stuff. So the same thing, if you run your own application inside the database, we can manage the database portion but we don't manage the application portion just like on the operating system. And so, there's still a management level that's required, no matter what, a level above that. And the other thing and I think this is what a lot of the stuff we're doing is based on is, you still have tons of stuff on-premises that needs full management. You have applications that you migrate that are not running Autonomous Linux, could be a Windows application that's running or it could be something on a different Linux distribution or you could still have some databases installed that you manage yourself, you don't want to use the autonomous or you're on a third-party. And so we want to make sure that we can address all of them with a single set of tools, right. >> Okay, so I wonder, can you give us just an overview, just briefly of the products that comprise into the cloud services, your management solution, what's in that portfolio? How should we think about it? >> Yeah, so it basically starts with Enterprise Manager on-premises, right? Which has been the tool that our Oracle database customers in particular have been using for many years and is widely used by our customer base. And so you have those customers, most of their real estate is on-premises and they can use enterprise management with local. They have it running and they don't want to change. They can keep doing that and we keep enhancing as you know, with newer versions of Enterprise Manager getting better. So, then there's the transition to cloud and so what we've been doing over the last several years is basically, looking at the things, well, one aspect is looking at things people, likes of Enterprise Manager and make sure that we provide similar functionality in Oracle cloud. So, we have Performance Hub for looking at how the database performance is working. We have APM for Application Performance Monitoring, we have Logging Analytics that looks at all the different log files and helps make sense of it for you. We have Database Management. So, a lot of the functionality that people like in Enterprise Manager mentioned the database that we've built into Oracle cloud, and, you know, a number of other things that are coming Operations Insights, to look at how databases are performing and how we can potentially do consolidation and stuff. So we've basically looked at what people have been using on-premises, how we can replicate that in Oracle cloud and then also, when you're in a cloud, how you can make make use of all the base services that a cloud vendor provides, telemetry, logging and so forth. And so, it's a broad portfolio and what it allows us to do with our customers is say, "Look, if you're predominantly on-prem, you want to stay there, keep using Enterprise Manager. If you're starting to move to Oracle cloud, you can first use EM, look at what's happening in the cloud and then switch over, start using all the management products we have in the cloud and let go of the Enterprise Manager instance on-premise. So you can gradually shift, you can start using more and more. Maybe you start with analytics first and then you start with insights and then you switch to database management. So there's a whole suite of possibilities. >> (indistinct) you mentioned APM, I've been watching that space, it's really evolved. I mean, you saw, you know, years ago, Splunk came out with sort of log analytics, maybe simplified that a little bit, now you're seeing some open source stuff come out. You're seeing a lot of startups come out, you saw Cisco made an acquisition with AppD and that whole space is transforming it seems that the future is all about that end to end visibility, simplifying the ability to remediate problems. And I'm thinking, okay, you just mentioned, you guys have a lot of these capabilities, you got Autonomous, is that sort of where you're headed with your capabilities? >> It definitely is and in fact, one of the... So, you know, APM allows you to say, "Hey, here's my web browser and it's making a connection to the database, to a middle tier" and it's hard for operations people in companies to say, hey, the end user calls and says, "You know, my order entry system is slow. Is it the browser? Is it the middle tier that they connect to? Is it the database that's overloaded in the backend?" And so, APM helps you with tracing, you know, what happens from where to where, where the delays are. Now, once you know where the delay is, you need to drill down on it. And then you need to go look at log files. And that's where the logging piece comes in. And what happens very often is that these log files are very difficult to read. You have networking log files and you have database log files and you have reslog files and you almost have to be an expert in all of these things. And so, then with Logging Analytics, we basically provide sort of an expert dashboard system on top of that, that allows us to say, "Hey! When you look at logging for the network stack, here are the most important errors that we could find." So you don't have to go and learn all the details of these things. And so, the real advantages of saying, "Hey, we have APM, we have Logging Analytics, we can tie the two together." Right, and so we can provide a solution that actually helps solve the problem, rather than, you need to use APM for one vendor, you need to use Logging Analytics from another vendor and you know, that doesn't necessarily work very well. >> Yeah and that's why you're seeing with like the ELK Stack it's cool, you're an open source guy, it's cool as an open source, but it's complicated to set up all that that brings. So, that's kind of a cool approach that you guys are taking. You mentioned Enterprise Manager, you just made a recent announcement, a new release. What's new in that new release? >> So Enterprise Manager 13.5 just got released. And so EM keeps improving, right? We've made a lot of changes over over the years and one of the things we've done in recent years is do more frequent updates sort of the cloud model frequent updates that are not just bug fixes but also introduce new functionality so people get more stuff more frequently rather than you know, once a year. And that's certainly been very attractive because it shows that it's a lively evolving product. And one of the main focus areas of course is cloud. And so a lot of work that happens in Enterprise Manager is hybrid cloud, which basically means I run Enterprise Manager and I have some stuff in Oracle cloud, I might have some other stuff in another cloud vendors environment and so we can actually see which databases are where and provide you with one consolidated view and one tool, right? And of course it supports Autonomous Database and Exadata in cloud servers and so forth. So you can from EM see both your databases on-premises and also how it's doing in in Oracle cloud as you potentially migrate things over. So that's one aspect. And then the other one is in terms of operations and automation. One of the things that we started doing again with Enterprise Manager in the last few years is making sure that everything has a REST API. So we try to make the experience with Enterprise Manager be very similar to how people work with a cloud service. Most folks now writing automation tools are used to calling REST APIs. EM in the early days didn't have REST APIs, now we're making sure everything works that way. And one of the advantages is that we can do extensibility without having to rewrite the product, that we just add the API clause in the agent and it makes it a lot easier to become part of the modern system. Another thing that we introduced last year but that we're evolving with more dashboards and so forth is the Grafana plugin. So even though Enterprise Manager provides lots of cool tools, a lot of cloud operations folks use a tool called Grafana. And so we provide a plugin that allows customers to have Grafana dashboards but the data actually comes out of Enterprise Manager. So that allows us to integrate EM into a more cloudy world in a cloud environment. I think the other important part is making sure that again, Enterprise Manager has sort of a cloud feel to it. So when you do patching and upgrades, it's near zero downtime which basically means that we do all the upgrades for you without having to bring EM down. Because even though it's a management tool, it's used for operations. So if there were downtime for patching Enterprise Manager for an hour, then for that hour, it's a blackout window for all the monitoring we do. And so we want to avoid that from happening, so now EM is upgrading, even though all the events are still happening and being processed, and then we do a very short switch. So that help our operations people to be more available. >> Yes. I mean, I've been talking about Automated Operations since, you know, lights out data centers since the eighties back in (laughs). I remember (indistinct) data center one-time lights out there were storage tech libraries in there and so... But there were a lot of unintended consequences around, you know, automated ops, and so people were sort of scared to go there, at least lean in too much but now with all this machine intelligence... So you're talking about ops automation, you mentioned the REST APIs, the Grafana plugins, the Cloud feel, is that what you're bringing to the table that's unique, is that unique to Oracle? >> Well, the integration with Oracle in that sense is unique. So one example is you mentioned the word migration, right? And so database migration tends to be something, you know, customers obviously take very serious. We go from one place, you have to move all your data to another place that runs in a slightly different environment. And so how do you know whether that migration is going to work? And you can't migrate a thousand databases manually, right? So automation, again, it's not just... Automation is not just to say, "Hey, I can do an upgrade of a system or I can make sure that nothing is done by hand when you patch something." It's more about having a huge fleet of servers and a huge fleet of databases. How can you move something from one place to another and automate that? And so with EM, you know, we start with sort of the prerequisite phase. So we're looking at the existing environment, how much memory does it need? How much storage does it use? Which version of the database does it have? How much data is there to move? Then on the target side, we see whether the target can actually run in that environment. Then we go and look at, you know, how do you want to migrate? Do you want to migrate everything from a sort of a physical model or do you want to migrate it from a logical model? Do you want to do it while your environment is still running so that you start backing up the data to the target database while your existing production system is still running? Then we do a short switch afterwards, or you say, "No, I want to bring my database down. I want to do the migrate and then bring it back up." So there's different deployment models that we can let our customers pick. And then when the migration is done, we have a ton of health checks that can validate whether the target database will run through basically the exact same way. And then you can say, "I want to migrate 10 databases or 50 databases" and it'll work, It's all automated out of the box. >> So you're saying, I mean, you've looked at the prevailing way you've done migrations, historically you'd have to freeze the code and then migrate, and it would take forever, it was a function of the number of lines of code you had. And then a lot of times, you know, people would say, "We're not going to freeze the code" and then they would almost go out of business trying to merge the two. You're saying in 2021, you can give customers the choice, you can migrate, you could change the, you know, refuel the plane while you're in midair? Is that essentially what you're saying? >> That's a good way of describing it, yeah. So your existing database is running and we can do a logical backup and restore. So while transactions are happening we're still migrating it over and then you can do a cutoff. It makes the transition a lot easier. But the other thing is that in the past, migrations would typically be two things. One is one database version to the next, more upgrades than migration. Then the second one is that old hardware or a different CPU architecture are moving to newer hardware in a new CPU architecture. Those were sort of the typical migrations that you had prior to Cloud. And from a CIS admin point of view or a DBA it was all something you could touch, that you could physically touch the boxes. When you move to cloud, it's this nebulous thing somewhere in a data center that you have no access to. And that by itself creates a barrier to a lot of admins and DBA's from saying, "Oh, it'll be okay." There's a lot of concern. And so by baking in all these tests and the prerequisites and all the dashboards to say, you know, "This is what you use. These are the features you use. We know that they're available on the other side so you can do the migration." It helps solve some of these problems and remove the barriers. >> Well that was just kind of same same vision when you guys came up with it. I don't know, quite a while ago now. And it took a while to get there with, you know, you had gen one and then gen two but that is, I think, unique to Oracle. I know maybe some others that are trying to do that as well, but you were really the first to do that and so... I want to switch topics to talk about security. It's hot topic. You guys, you know, like many companies really focused on security. Does Enterprise Manager bring any of that over? I mean, the prevailing way to do security often times is to do scripts and write, you know, custom security policy scripts are fragile, they break, what can you tell us about security? >> Yeah. So there's really two things, you know. One is, we obviously have our own best security practices. How we run a database inside Oracle for our own world, we've learned about that over the years. And so we sort of baked that knowledge into Enterprise Manager. So we can say, "Hey, if you install this way, we do the install and the configuration based on our best practice." That's one thing. The other one is there's STIG, there's PCI and they're ShipBob, those are the main ones. And so customers can do their own way. They can download the documentation and do it manually. But what we've done is, and we've done this for a long time, is basically bake those policies into Enterprise Manager. So you can say, "Here's my database this needs to be PCI compliant or it needs to be HIPAA compliant and you push a button and then we validate the policies in those documents or in those prescript described files. And we make sure that the database is combined to that. And so we take that manual work and all that stuff basically out of the picture, we say, "Push this button and we'll take care of it." >> Now, Wim, but just quick sidebar here, last time we talked, it was under a year ago. It was definitely during COVID and it's still during COVID. We talked about the state of the penguin. So I'm wondering, you know, what's the latest update for Linux, any Linux developments that we should be aware of? >> Linux, we're still working very hard on Autonomous Linux and that's something where we can really differentiate and solve a problem. Of course, one of the things to mention is that Enterprise Manager can can do HIPAA compliance on Oracle Linux as well. So the security practices are not just for the database it can also go down to the operating system. Anyway, so on the Autonomous Linux side, you know, management in an Oracle Cloud's OS management is evolving. We're spending a lot of time on integrating log capturing, and if something were to go wrong that we can analyze a log file on the fly and send you a notification saying, "Hey, you know there was this bug and here's the cause." And it was potentially a fix for it to Autonomous Linux and we're putting a lot of effort into that. And then also sort of IT/operation management where we can look at the different applications that are running. So you're running a web server on a Linux environment or you're running some Java processes, we can see what's running. We can say, "Hey, here's the CPU utilization over the past week or the past year." And then how is this evolving? Say, if something suddenly spikes we can say, "Well, that's normal, because every Monday morning at 10 o'clock there's a spike or this is abnormal." And then you can start drilling this down. And this comes back to overtime integration with whether it's APM or Logging Analytics, we can tie the dots, right? We can connect them, we can say, "Push this thing, then click on that link." We give you the information. So it's that integration with the entire cloud platform that's really happening now >> Integration, there's that theme again. I want to come back to migration and I think you did a good job of explaining how you sort of make that non-disruptive and you know, your customers, I think, you know, generally you're pushing you know, that experience which makes people more comfortable. But my question is, why do people want to migrate if it works and it's on prem, are they doing it just because they want to get out of the data center business? Or is it a better experience in the cloud? What can you tell us there? >> You know, it's a little bit of everything. You know, one is, of course the idea that data center maintenance costs are very high. The other one is that when you run your own data center, you know, we obviously have this problem but when you're a cloud vendor, you have these problems but we're in this business. But if you buy a server, then in three years that server basically is depreciated by new versions and they have to do migration stuff. And so one of the advantages with cloud is you push a button, you have a new version of the hardware, basically, right? So the refreshes happen on a regular basis. You don't have to go and recycle that yourself. Then the other part is the subscription model. It's a lot easier to pay for what you use rather than you have a data center whether it's used or not, you pay for it. So there's the cost advantages and predictability of what you need, you pay for, you can say, "Oh next year we need to get x more of EMs." And it's easier to scale that, right? We take care of dealing with capacity planning. You don't have to deal with capacity planning of hardware, we do that as the cloud vendor. So there's all these practical advantages you get from doing it remotely and that's really what the appeal is. >> Right. So, as it relates to Enterprise Manager, did you guys have to like tear down the code and rebuild it? Was it entire like redo? How did you achieve that? >> No, no, no. So, Enterprise Manager keeps evolving and you know, we changed the underlying technologies here and there, piecemeal, not sort of a wholesale replacement. And so in talking about five, there's a lot of new stuff but it's built on the existing EM core. And so we're just, you know, improving certain areas. One of the things is, stability is important for our customers, obviously. And so by picking things piecemeal, we replace one engine rather than the whole thing. It allows us to introduce change more slowly, right. And then it's well-tested as a unit and then when we go on to the next thing. And then the other one is I mentioned earlier, a lot of the automation and extensibility comes from REST APIs. And so instead of basically re-writing everything we just provide a REST endpoint and we make all the new features that we built automatically be REST enabled. So that makes it a lot easier for us to introduce new stuff. >> Got it. So if I want to poke around with this new version of Enterprise Manager, can I do that? Is there a place I can go, do I have to call a rep? How does that work? >> Yeah, so for information you can just go to oracle.com/enterprise manager. That's the website that has all the data. The other thing is if you're already playing with Oracle Cloud or you use Oracle Cloud, we have Enterprise Manager images in the marketplace. So if you have never used EM, you can go to Oracle Cloud, push a button in the marketplace and you get a full Enterprise Manager installation in a matter of minutes. And then you can just start using that as well. >> Awesome. Hey, I wanted to ask you about, you know, people forget that you guys are the stewards of MySQL and we've been looking at MySQL Database Cloud service with HeatWave Did you name that? And so I wonder if you could talk about what you're doing with regard to managing HeatWave environments? >> So, HeatWave is the MySQL option that helps with analytics, right? And it really accelerates MySQL usage by 100 x and in some cases more and it's transparent to the customer. So as a MySQL user, you connect with standard MySQL applications and APIs and SQL and everything. And the HeatWave part is all done within the MySQL server. The engine itself says, "Oh, this SQL query, we can offload to the backend HeatWave cluster," which then goes in memory operations and blazingly fast returns it to you. And so the nice thing is that it turns every single MySQL database into also a data warehouse without any change whatsoever in your application. So it's been widely popular and it's quite exciting. I didn't personally name it, HeatWave, that was not my decision, but it sounds very cool. >> That's very cool. >> Yeah, It's a very cool name. >> We love MySQL, we started our company on the lamp stack, so like many >> Oh? >> Yeah, yeah. >> Yeah, yeah. That's great. So, yeah. And so with HeatWave or MySQL in general we're basically doing the same thing as we have done for the Oracle Database. So we're going to add more functionality in our database management tools to also look at HeatWave. So whether it's doing things like performance hub or generic database management and monitoring tools, we'll expand that in, you know, in the near future, in the future. >> That's great. Well, Wim, it's always a pleasure. Thank you so much for coming back in "The Cube" and letting me ask all my Colombo questions. It was really a pleasure having you. (mumbling) >> It's good be here. Thank you so much. >> You're welcome. And thank you for watching, everybody, this is Dave Vellante. We'll see you next time. (bright music)

>> From theCUBE studios in Palo Alto and Boston, connecting with thought-leaders all around the world, this is a Cube Conversation. >> Hi everybody, this is Dave Vellante. Welcome to this Cube Conversation. We're really excited to have Wim Coekaerts in, he is the senior vice-president of software development at Oracle. Wim, it's great to have you on, and, you know I often say I wish we were face-to-face but if we were you'd have to cut off my tie, cause developers and ties just don't go together. >> No, I know, and this is my normal outfit, so this is me wherever I go. Hi again, good to see you. >> Yeah, great to see you. So, of course, you know a lot of people are confused about Oracle, and open-source, they say "Oracle? Open-source? What is that all about?" But I think you're misunderstood. People don't, first of all, realize you as the leader of the software-development community inside of Oracle, I mean, you've been involved in Linux since the early 90s. But you guys have a lot of committers, you do a lot. I want to talk about that. What is up with Oracle, and open-source? >> Ah, well, it's a broad question. So, you know, a couple of things. One is, we have many different areas within the company that are dealing with open-source. So we have the cloud team doing a lot of stuff around cloud SDKs and support for different languages like Python and Go, and of course Java and so forth, so they do a lot around ensuring that the Oracle ecosystem is integrated in the open-source tools that customers use, or developers use, Terraform companies and so forth. And then you have the Java team, and so forth. Java is open-source and then the Graal project, GraalVM which is a polyglot compiler that can run Java, and Python, and Javascript and so forth together in one. VM do really cool optimizations, that's an open-source project, also on GitHub. There's of course MySQL, which is along with Java, they're probably the two most popular and widely used open-source projects out there. There's VirtualBox which is of course also a very popular project that's open-source. There's all the work we do around Linux. And I think one of the things is that, when you have so many different areas, doing things that are for that area, then as a developer or as a customer, you typically just deal with that group. And what you see is, oh you're talking to the Java developers, so you know what's going on around Java. The Java developers might not necessarily say, "Oh well we also do MySQL, and we do Linux and VirtualBox and so forth," and so you get a rather myopic, narrow view of the larger company. When you add all these things up, and there will be one big slide that says "This is Oracle, these are all these open source projects," and there's multiple ways. One is, we have projects that we've open-sourced and all the code came from us and we made it publicly available, we're the main contributor and we get contributions back. There are other projects where we contribute to third-party in terms of enhancing things, like I said with the Cloud Team, and then in general something like Linux where we're part of an external project and we participate in development of that project at large. And so there's these three different ways, when you count up all the developers that we have that deal with open-source on a daily basis. And in terms of contributions, in terms of bug fixes, testing, and so forth, it's thousands, literally, full-time paid developers. And of course, all the projects are all either on GitHub or similar sites that are very popular. So yeah, I think the misunderstood is probably a lack of knowledge of the breadth of what we do. And, you know, our primary goal is to provide services and products to customers, and so the open-source part is sort of embedded in a development methodology. But that's not something we sell or market separately, we just work with customers and products and services, and so in some cases it's not well-understood. >> Yeah. Well, we're talking of course, we're talking about the state of the penguin, I think it's important for people to understand, Oracle got into the Linux game in the 90s, maybe the latter part of the 90s and Oracle, of course, wants to make Linux-- wants to make Oracle, it's applications and database run better on Linux, but as you're pointing out, your Linux distro, full support, end-to-end, thousands of people in your open-source community, and the contributions that you make to Linux, many if not most, they go upstream, everybody can benefit from those, but of course you want an Oracle distro that is going to make Oracle stuff run better, that's always kind of been the Oracle way. >> Well, so, yes, two things though. One is, so everything we do is upstream. So we have no Linux patches that are not contributed upstream; There's no proprietary code in Oracle Linux at all, it's all completely open, publicly available: the source code, the change log, all the commits, it's fully open and public, which sometimes is not well-understood, but it's completely open. And, everything we do in terms of feature development or functionality or bug fixes goes upstream to the Linux kernel mail-list. It's actually, it's the only way to be able to manage a Linux distribution and be a Linux vendor is to live in that eco-system. Otherwise, the cost of maintaining your own fork, so to speak, is very high and it doesn't really solve the problem. Now, the functionality we work on obviously is focused on making Oracle products run better, making Oracle Cloud run better, and so forth. However, again, what's important to understand, though, is an Oracle database is a program running on an operating system. It does IO, it does networking, it deals with memory management, lots of processing. So, for the most part, the things that we work on to improve that helps everyone out, right? It helps every other database run better, or helps every other language run better. So none of these changes are specific to Oracle, they're just things that we found doing performance benchmarks and testing and so forth, where we say "Hey, if Linux did the following, it would make boot-up faster. Now boot-up has nothing to do with the database. But our customers run on 1-terabyte, 4-terabyte, 8-terabyte systems, and so booting up, and Linux starting up, and cleaning up memory takes a long time. So we want to reduce that from an availability point of view. So here, we're now talking about just enterprise for you. So there's this broad set of things we work on that definitely help us, but they're actually really completely generic and help everyone out. >> Yeah, that's great. So I wanted to kind of get that out of the way and help our audience understand that. So let's get into it a little bit; What are you seeing, what's going on in IT, pick your observation space and your vision of what you see happening out there. >> Well, you know, it's very interesting, it's sort of, there's two... there's sort of two worlds, right, there's the cloud world and the move to cloud, and there's the on-premises world, where people run their systems on their own. And, one of the things that we've learned is, when you talk about machine-learning, obviously, is something that's very popular these days, and automation. And so in order to rely on machine-learning well, and have algorithms that are very effective, you need lots of data. And so being a cloud vendor, and having Linux in our cloud on tens of thousands, or hundreds of thousands of servers, or more, allows us to have a view of how an operating system works across an incredibly large scale. So we get lots of data. And so for us to figure out which algorithms work well in terms of how can we do network optimizations, how can we discover anomalies on the storage site, and deal with it and so forth, we can do that at scale. And what's interesting is, how do we then bring that on-prem? Well, if we can get the data and the learning done, the training done, in our cloud directly, then when we provide that service also for people running Oracle Linux on premises then that will work. The alternative is to have point solutions where you provide something to a customer, and he needs to learn something from small amounts of data. That doesn't work so well. So I think having both worlds, on-prem and cloud directly, allows us to kind of benefit from that. And I think that's important, because lots of customers are interested in going to cloud. Many of the enterprises have not yet. You know, they're starting, but there's still a huge on-premises space that's important. And so by being able to get them familiar with how these things work at scale, autonomy is again important, right, Autonomous Database is incredibly popular and so forth, that allows us to then say, "Here, try these things out here, it's a service. We can show you the benefits right away," and then as that improves we bring that, to a certain extent, on-premises as well. And then they can have it in both places. And that, I think, is something, again, that's relatively unique but also very important, is that we want to provide services and products that act similarly on-premises as well as in cloud, because at some point when people move we want to make that transition seamless. And what you have today for the most part is one world that's on-prem, and then the cloud world is completely different. And that is a big barrier of moving, and so we want to reduce that, we can run the same operating system local as well as cloud, you can the same functionality, and then that helps transition people over much easier. >> Yeah, well Oracle actually was one of the -- I think Oracle was the first company to actually market same-same, you actually used that term. Others put forth that concept, but Oracle was the first to announce products like Cloud at Customer, that were same-same, now it took some time to actually get it perfected, and get it to market, but the point is, and we've written about this, is Oracle, because of the ascendancy of cloud, flipped and has a cloud-first mentality, and you just kind of referenced that, you just said, "And you can bring that to on-prem." So I wonder if you could talk about that cloud-first mentality, and the impact on hybrid. >> So yeah, I think the cloud-first part is of course in cloud we work on services moreso than products that we deliver. And there's a number of things that are happening. So one is that we obviously continue to provide products to customers, you can download Oracle Linux, you can download the database and what not, you can install it on your own, you can do the traditional way of working. Then in the cloud-world, what typically happens is "Oh, I use a database service. I'm not installing anything, I push a button and I get an IP address and a SQL that connects extremely quickly to the database." And we take care of everything underneath that is on this database. Now, in order to do that, you need a whole infrastructure in place, you need log-in agents, you need a back-end that captures all that stuff, you need monitoring tools, you need all the automation scripts for bringing the service up and monitor it. And so, that takes a lot of time to do right, and we learn a lot by doing this. And so the cloud-first part of these services means that we get to experience this ourselves with direct access to everything. Now taking that service with all of the additional features like autonomy, and bringing that to an on-premises world, we have to make sure we can package that so that all these pieces around it go along with it. And that takes a little bit more time, so we can do everything at the same time. And so what we've done with Autonomous Database is we created everything in Oracle Cloud, we have the whole system running really well, and then we've been able to sort of package that and shrink it into something that can be installed on-premises, but then connected into Oracle Cloud again. And so that way we can get all the telemetry over the metric, and that allows us to scale. Because part of providing a cloud service that runs on-prem in the customer environment is that we need to be able to remotely manage that similar to how that runs in our own cloud. Right, otherwise it doesn't scale. And so that takes a little bit of time, but we've done all that work, and now with Cloud at Customer Database that's really in place. >> Yeah, you really want to have that same cloud experience, whether with on-prem, in the public cloud, hybrid, et cetera. So, I want to explore a little bit more who is using Oracle Linux, and what's the driver for using it. Can you describe maybe some of the types of customers and why they buy? >> Sure, so we started this fourteen years ago, in 2006, October 25th, 2006. I remember that day very well; Penguins on stage and a big launch for Oracle Linux in San Francisco Moscone Center. So, look, the initial driver for Oracle Linux was to ensure that Oracle database customers or Oracle product customers had a good operating system experience, and the ability to be able to handle critical issues when that occurs, because typically a database runs the company's critical data: the most essential stuff that a company has is typically in a database, an Oracle database. And so when that thing has issues with the operating system, you don't want just to talk to multiple vendors and have finger-pointing, and having to explain to an operating system vendor how the database works. In the Unix world, we had a good relationship with the OS vendors, and the hardware vendors, they were the same. And they knew our products really well, and in the Linux world, that was very different. The OS vendor basically did not want to understand or learn anything about the products living on top. And so while to a certain extent that makes sense, it's an enterprise world where time is of the essence, and downtime needs to be limited absolutely. We can't have these arguments and such. And that was the driver, initially, for doing Oracle Linux. It was to ensure there was a Linux distribution really backed by us, that we could fix, that we could fully support. That was completely the original intent. And so the early customer base was database customers. Database and middleware. Mostly database. But that has then evolved quickly, and so what happened was, people say "Look, I have a thousand servers, a hundred run Oracle, so we'll run Oracle Linux on those hundred, and we'll run something else on those other nine-hundred." Now after a year or so, they realize that our support is really good; We fix all these issues, and so then they're like "Why are we having two Linux distributions? This thing works really well, it runs any application, it's fully compatible, so we'll do a thousand with Oracle Linux." And so the early days, the first few years, was definitely Oracle Database as the core driver, and then it sort of expanded to the rest of the estate. And over the years, we've added lots of features and functionality, like Ksplice, and so forth. We have an attractive pricing model for running on servers, and so now lots of our customers have a very small Oracle percentage running and many other things running. So it's really become a all-or-nothing play in the Linux space, and we're well-known now, so it's actually very good. >> You just mentioned Ksplice. We've been talking about cloud, and on-prem, and hybrid. Let's talk about security, because security really is a differentiator, particularly if you're going to start to put stuff in the cloud. Talk about Ksplice specifically, but generally security and your policy there. >> So, "Security first" is sort of what you hear us say and do, in everything we do. The database obviously security, on the Linux site security matters. Ksplice as a technology is there to do critical bug-fixing and make sure that we can apply security vulnerability fixes without affecting the customer, and not have downtime. And if you look at most of the cases or many of the cases where you have security vulnerabilities and exploits, it tends to be because systems were not patched. Why were they not patched? Well not that our customer doesn't understand that it's important, but it's a whole train of events that needs to happen. You have to, you get notified that there's a security issue in your operating system or application. Then, well, an application typically means it's a multi-layered setup. So if you have to bring your database server down, then you first have to coordinate with the application users to bring the app server down, cause that talks to the database. So to patch one system, you basically have to bring down the whole application stack. You have to negotiate with the DBAs, you have to negotiate with the app admins, you have to negotiate with the user. It takes weeks to do that and find time. Well during that time, you're vulnerable. So the only way, really, to address security in a scalable and reducing that window of time is to do it without affecting the customer. And so Casewise is something that, it's a company we acquired in 2009, and have since evolved in terms of capabilities, and so it allows us to patch the Linux terminal without downtime. We lock the kernel for 8 microseconds. It's literally no downtime. You don't have to bring down applications, the user doesn't see it, there's no hang, there's no delay. And so by doing that, you can run a Linux operating system, or gLinux, and you can be fully patched on a system that hasn't rebooted for 3 years. You don't even know it. And so by doing that type of stuff, it makes customers more secure, and it avoids them-- It saves them a lot of money in terms of dealing with project management and so forth, but it really keeps them secure. And so we do that for the Linux kernel, we do that for some of the libraries on top that are critical like OpenSSL and 2 LVC, and, you know one example-- I can give you two examples. So one example is, Heartbleed was this bug in OpenSSL a number of years ago. And so everyone had to patch their SSH server. And that meant, basically, systems around the world had to reboot. Like a whole IT reboot across the world. With Ksplice today, if Heartbleed were to happen tomorrow, we would be able to patch this online for all the Oracle Linux customers without any downtime. No reboots, no restarting of applications, everything keeps running. The amount of money saved would be massive, and also, of course, the headache. Another example is, and this was in Oracle Cloud, when some of these CPU bugs that happened a few years ago that were rather damaging on the cloud side, where you could basically see memory potentially of other CPUs running, the cloud is incredibly critical. We were basically able to basically patch our entire cloud in four hours. And the customer didn't know, right, a hundred and twenty million patches, or something, that we applied within four hours, all online, without any downtime. And so that technology has been really helpful, both for us to run our cloud, but the exact same patches and same fixes go to customers on-premises as well. But this comes back to the whole, what we do in cloud we also do for customer. And I think that's a unique thing that we have at Oracle which is quite fascinating. The operating system we run for our customers, the operating system that's the host part of VMs, is the exact same binary and source code that we make available, just to be clear, the exact same binaries are the ones that you run as a customer on-premises. So if you run Oracle Linux with KVM, you run VMs, you're actually running the exact same stuff as we run underneath our customer's stuff. Nobody else does that, everyone else has a black box. So I think that helps a little bit with transparency as well. >> Yeah, and that homogeneity just creates an environment, you're talking about that sort of security mindset, it's critical, you're not just bolting it on, it's part of the culture. But you started your career, and then of course you were a Linux person when you came to Oracle, but then I think you spent some time in database, back in the day when there were serious database wars going on, before Oracle became the king of database. So now you've got, obviously, this great portfolio, and a lot of really sharp software developers; What should we expect going forward, from Oracle? What should we look for? >> You know, I was talking to some, I was welcoming some interns to the company, for their summer internship yesterday, and one of the things I mentioned to them was that -- so cloud obviously gives us a lot of opportunities, but there's a number of things. One is, we have such a breadth of applications and software and hardware together. We have the servers, we have the storage, we have the operating systems, we have the database layer and so forth, and we have the cloud side, and one of the great opportunities, and I think we've shown a lot of this happening with the ability to create something like Autonomous Database, is to combine all these things. Right, we have such a broad portfolio of really cool technology that by itself is okay, but if you combine the things it really becomes awesome. You cannot create autonomous database without having autonomous learning. You cannot create those two and make them really safe without also controlling the firmware on the hardware and so forth. So by being able to combine all these layers, and by having a really great relationship across the teams within the company, that opens up a lot of opportunities to do stuff really quickly. And having the scale for that. I think that has been, for the last few years, a really great thing, but I can see that being one of the advantages that we have going forward. We have Oracle Fusion Applications, which is incredibly popular, and has great growth, and then we have that running on Oracle Cloud, that talks to Oracle Autonomous Database, so we bring all these pieces together. And no other SaaS vendor can do that, because they don't have these other pieces. They have one area, we have all of them. And so that's the exciting part for me, it's not so much about making my own world better, and having Linux be better, and Casewise and so forth, which is important, but that becoming part of the bigger picture. And that's the exciting part. >> Well, Oracle's always invested in RND, we've made that point many, many times. Whether it's database, you know Fusion was a painful but worthy effort, the whole public cloud piece, obviously many acquisitions, but the investments that you've made in open-source as well, Wim, you're a great spokesperson, and a great representative of the open-source community generally, and then Oracle specifically, so thanks very much for coming on theCUBE and sharing with us the state of the penguin, and best of luck. >> You're welcome. Thank you, thanks for having me. >> Alright, and thank you for watching, everybody. This is Dave Vellante for theCUBE. We'll see you next time. (cheerful music).

>> From theCUBE studios in Palo Alto and Boston, connecting with thought-leaders all around the world, this is a Cube Conversation. >> Hi everybody, this is Dave Vellante. Welcome to this Cube Conversation. We're really excited to have Wim Coekaerts in, he is the senior vice-president of software development at Oracle. Wim, it's great to have you on, and, you know I often say I wish we were face-to-face but if we were you'd have to cut off my tie, cause developers and ties just don't go together. >> No, I know, and this is my normal outfit, so this is me wherever I go. Hi again, good to see you. >> Yeah, great to see you. So, of course, you know a lot of people are confused about Oracle, and open-source, they say "Oracle? Open-source? What is that all about?" But I think you're misunderstood. People don't, first of all, realize you as the leader of the software-development community inside of Oracle, I mean, you've been involved in Linux since the early 90s. But you guys have a lot of committers, you do a lot. I want to talk about that. What is up with Oracle, and open-source? >> Ah, well, it's a broad question. So, you know, a couple of things. One is, we have many different areas within the company that are dealing with open-source. So we have the cloud team doing a lot of stuff around cloud SDKs and support for different languages like Python and Go, and of course Java and so forth, so they do a lot around ensuring that the Oracle ecosystem is integrated in the open-source tools that customers use, or developers use, Terraform companies and so forth. And then you have the Java team, and so forth. Java is open-source and then the Graal project, GraalVM which is a polyglot compiler that can run Java, and Python, and Javascript and so forth together in one. VM do really cool optimizations, that's an open-source project, also on GitHub. There's of course MySQL, which is along with Java, they're probably the two most popular and widely used open-source projects out there. There's VirtualBox which is of course also a very popular project that's open-source. There's all the work we do around Linux. And I think one of the things is that, when you have so many different areas, doing things that are for that area, then as a developer or as a customer, you typically just deal with that group. And what you see is, oh you're talking to the Java developers, so you know what's going on around Java. The Java developers might not necessarily say, "Oh well we also do MySQL, and we do Linux and VirtualBox and so forth," and so you get a rather myopic, narrow view of the larger company. When you add all these things up, and there will be one big slide that says "This is Oracle, these are all these open source projects," and there's multiple ways. One is, we have projects that we've open-sourced and all the code came from us and we made it publicly available, we're the main contributor and we get contributions back. There are other projects where we contribute to third-party in terms of enhancing things, like I said with the Cloud Team, and then in general something like Linux where we're part of an external project and we participate in development of that project at large. And so there's these three different ways, when you count up all the developers that we have that deal with open-source on a daily basis. And in terms of contributions, in terms of bug fixes, testing, and so forth, it's thousands, literally, full-time paid developers. And of course, all the projects are all either on GitHub or similar sites that are very popular. So yeah, I think the misunderstood is probably a lack of knowledge of the breadth of what we do. And, you know, our primary goal is to provide services and products to customers, and so the open-source part is sort of embedded in a development methodology. But that's not something we sell or market separately, we just work with customers and products and services, and so in some cases it's not well-understood. >> Yeah. Well, we're talking of course, we're talking about the state of the penguin, I think it's important for people to understand, Oracle got into the Linux game in the 90s, maybe the latter part of the 90s and Oracle, of course, wants to make Linux-- wants to make Oracle, it's applications and database run better on Linux, but as you're pointing out, your Linux distro, full support, end-to-end, thousands of people in your open-source community, and the contributions that you make to Linux, many if not most, they go upstream, everybody can benefit from those, but of course you want an Oracle distro that is going to make Oracle stuff run better, that's always kind of been the Oracle way. >> Well, so, yes, two things though. One is, so everything we do is upstream. So we have no Linux patches that are not contributed upstream; There's no proprietary code in Oracle Linux at all, it's all completely open, publicly available: the source code, the change log, all the commits, it's fully open and public, which sometimes is not well-understood, but it's completely open. And, everything we do in terms of feature development or functionality or bug fixes goes upstream to the Linux kernel mail-list. It's actually, it's the only way to be able to manage a Linux distribution and be a Linux vendor is to live in that eco-system. Otherwise, the cost of maintaining your own fork, so to speak, is very high and it doesn't really solve the problem. Now, the functionality we work on obviously is focused on making Oracle products run better, making Oracle Cloud run better, and so forth. However, again, what's important to understand, though, is an Oracle database is a program running on an operating system. It does IO, it does networking, it deals with memory management, lots of processing. So, for the most part, the things that we work on to improve that helps everyone out, right? It helps every other database run better, or helps every other language run better. So none of these changes are specific to Oracle, they're just things that we found doing performance benchmarks and testing and so forth, where we say "Hey, if Linux did the following, it would make boot-up faster. Now boot-up has nothing to do with the database. But our customers run on 1-terabyte, 4-terabyte, 8-terabyte systems, and so booting up, and Linux starting up, and cleaning up memory takes a long time. So we want to reduce that from an availability point of view. So here, we're now talking about just enterprise for you. So there's this broad set of things we work on that definitely help us, but they're actually really completely generic and help everyone out. >> Yeah, that's great. So I wanted to kind of get that out of the way and help our audience understand that. So let's get into it a little bit; What are you seeing, what's going on in IT, pick your observation space and your vision of what you see happening out there. >> Well, you know, it's very interesting, it's sort of, there's two... there's sort of two worlds, right, there's the cloud world and the move to cloud, and there's the on-premises world, where people run their systems on their own. And, one of the things that we've learned is, when you talk about machine-learning, obviously, is something that's very popular these days, and automation. And so in order to rely on machine-learning well, and have algorithms that are very effective, you need lots of data. And so being a cloud vendor, and having Linux in our cloud on tens of thousands, or hundreds of thousands of servers, or more, allows us to have a view of how an operating system works across an incredibly large scale. So we get lots of data. And so for us to figure out which algorithms work well in terms of how can we do network optimizations, how can we discover anomalies on the storage site, and deal with it and so forth, we can do that at scale. And what's interesting is, how do we then bring that on-prem? Well, if we can get the data and the learning done, the training done, in our cloud directly, then when we provide that service also for people running Oracle Linux on premises then that will work. The alternative is to have point solutions where you provide something to a customer, and he needs to learn something from small amounts of data. That doesn't work so well. So I think having both worlds, on-prem and cloud directly, allows us to kind of benefit from that. And I think that's important, because lots of customers are interested in going to cloud. Many of the enterprises have not yet. You know, they're starting, but there's still a huge on-premises space that's important. And so by being able to get them familiar with how these things work at scale, autonomy is again important, right, Autonomous Database is incredibly popular and so forth, that allows us to then say, "Here, try these things out here, it's a service. We can show you the benefits right away," and then as that improves we bring that, to a certain extent, on-premises as well. And then they can have it in both places. And that, I think, is something, again, that's relatively unique but also very important, is that we want to provide services and products that act similarly on-premises as well as in cloud, because at some point when people move we want to make that transition seamless. And what you have today for the most part is one world that's on-prem, and then the cloud world is completely different. And that is a big barrier of moving, and so we want to reduce that, we can run the same operating system local as well as cloud, you can the same functionality, and then that helps transition people over much easier. >> Yeah, well Oracle actually was one of the -- I think Oracle was the first company to actually market same-same, you actually used that term. Others put forth that concept, but Oracle was the first to announce products like Cloud at Customer, that were same-same, now it took some time to actually get it perfected, and get it to market, but the point is, and we've written about this, is Oracle, because of the ascendancy of cloud, flipped and has a cloud-first mentality, and you just kind of referenced that, you just said, "And you can bring that to on-prem." So I wonder if you could talk about that cloud-first mentality, and the impact on hybrid. >> So yeah, I think the cloud-first part is of course in cloud we work on services moreso than products that we deliver. And there's a number of things that are happening. So one is that we obviously continue to provide products to customers, you can download Oracle Linux, you can download the database and what not, you can install it on your own, you can do the traditional way of working. Then in the cloud-world, what typically happens is "Oh, I use a database service. I'm not installing anything, I push a button and I get an IP address and a SQL that connects extremely quickly to the database." And we take care of everything underneath that is on this database. Now, in order to do that, you need a whole infrastructure in place, you need log-in agents, you need a back-end that captures all that stuff, you need monitoring tools, you need all the automation scripts for bringing the service up and monitor it. And so, that takes a lot of time to do right, and we learn a lot by doing this. And so the cloud-first part of these services means that we get to experience this ourselves with direct access to everything. Now taking that service with all of the additional features like autonomy, and bringing that to an on-premises world, we have to make sure we can package that so that all these pieces around it go along with it. And that takes a little bit more time, so we can do everything at the same time. And so what we've done with Autonomous Database is we created everything in Oracle Cloud, we have the whole system running really well, and then we've been able to sort of package that and shrink it into something that can be installed on-premises, but then connected into Oracle Cloud again. And so that way we can get all the telemetry over the metric, and that allows us to scale. Because part of providing a cloud service that runs on-prem in the customer environment is that we need to be able to remotely manage that similar to how that runs in our own cloud. Right, otherwise it doesn't scale. And so that takes a little bit of time, but we've done all that work, and now with Cloud at Customer Database that's really in place. >> Yeah, you really want to have that same cloud experience, whether with on-prem, in the public cloud, hybrid, et cetera. So, I want to explore a little bit more who is using Oracle Linux, and what's the driver for using it. Can you describe maybe some of the types of customers and why they buy? >> Sure, so we started this fourteen years ago, in 2006, October 25th, 2006. I remember that day very well; Penguins on stage and a big launch for Oracle Linux in San Francisco Moscone Center. So, look, the initial driver for Oracle Linux was to ensure that Oracle database customers or Oracle product customers had a good operating system experience, and the ability to be able to handle critical issues when that occurs, because typically a database runs the company's critical data: the most essential stuff that a company has is typically in a database, an Oracle database. And so when that thing has issues with the operating system, you don't want just to talk to multiple vendors and have finger-pointing, and having to explain to an operating system vendor how the database works. In the Unix world, we had a good relationship with the OS vendors, and the hardware vendors, they were the same. And they knew our products really well, and in the Linux world, that was very different. The OS vendor basically did not want to understand or learn anything about the products living on top. And so while to a certain extent that makes sense, it's an enterprise world where time is of the essence, and downtime needs to be limited absolutely. We can't have these arguments and such. And that was the driver, initially, for doing Oracle Linux. It was to ensure there was a Linux distribution really backed by us, that we could fix, that we could fully support. That was completely the original intent. And so the early customer base was database customers. Database and middleware. Mostly database. But that has then evolved quickly, and so what happened was, people say "Look, I have a thousand servers, a hundred run Oracle, so we'll run Oracle Linux on those hundred, and we'll run something else on those other nine-hundred." Now after a year or so, they realize that our support is really good; We fix all these issues, and so then they're like "Why are we having two Linux distributions? This thing works really well, it runs any application, it's fully compatible, so we'll do a thousand with Oracle Linux." And so the early days, the first few years, was definitely Oracle Database as the core driver, and then it sort of expanded to the rest of the estate. And over the years, we've added lots of features and functionality, like Ksplice, and so forth. We have an attractive pricing model for running on servers, and so now lots of our customers have a very small Oracle percentage running and many other things running. So it's really become a all-or-nothing play in the Linux space, and we're well-known now, so it's actually very good. >> You just mentioned Ksplice. We've been talking about cloud, and on-prem, and hybrid. Let's talk about security, because security really is a differentiator, particularly if you're going to start to put stuff in the cloud. Talk about Ksplice specifically, but generally security and your policy there. >> So, "Security first" is sort of what you hear us say and do, in everything we do. The database obviously security, on the Linux site security matters. Ksplice as a technology is there to do critical bug-fixing and make sure that we can apply security vulnerability fixes without affecting the customer, and not have downtime. And if you look at most of the cases or many of the cases where you have security vulnerabilities and exploits, it tends to be because systems were not patched. Why were they not patched? Well not that our customer doesn't understand that it's important, but it's a whole train of events that needs to happen. You have to, you get notified that there's a security issue in your operating system or application. Then, well, an application typically means it's a multi-layered setup. So if you have to bring your database server down, then you first have to coordinate with the application users to bring the app server down, cause that talks to the database. So to patch one system, you basically have to bring down the whole application stack. You have to negotiate with the DBAs, you have to negotiate with the app admins, you have to negotiate with the user. It takes weeks to do that and find time. Well during that time, you're vulnerable. So the only way, really, to address security in a scalable and reducing that window of time is to do it without affecting the customer. And so Casewise is something that, it's a company we acquired in 2009, and have since evolved in terms of capabilities, and so it allows us to patch the Linux terminal without downtime. We lock the kernel for 8 microseconds. It's literally no downtime. You don't have to bring down applications, the user doesn't see it, there's no hang, there's no delay. And so by doing that, you can run a Linux operating system, or gLinux, and you can be fully patched on a system that hasn't rebooted for 3 years. You don't even know it. And so by doing that type of stuff, it makes customers more secure, and it avoids them-- It saves them a lot of money in terms of dealing with project management and so forth, but it really keeps them secure. And so we do that for the Linux kernel, we do that for some of the libraries on top that are critical like OpenSSL and 2 LVC, and, you know one example-- I can give you two examples. So one example is, Heartbleed was this bug in OpenSSL a number of years ago. And so everyone had to patch their SSH server. And that meant, basically, systems around the world had to reboot. Like a whole IT reboot across the world. With Ksplice today, if Heartbleed were to happen tomorrow, we would be able to patch this online for all the Oracle Linux customers without any downtime. No reboots, no restarting of applications, everything keeps running. The amount of money saved would be massive, and also, of course, the headache. Another example is, and this was in Oracle Cloud, when some of these CPU bugs that happened a few years ago that were rather damaging on the cloud side, where you could basically see memory potentially of other CPUs running, the cloud is incredibly critical. We were basically able to basically patch our entire cloud in four hours. And the customer didn't know, right, a hundred and twenty million patches, or something, that we applied within four hours, all online, without any downtime. And so that technology has been really helpful, both for us to run our cloud, but the exact same patches and same fixes go to customers on-premises as well. But this comes back to the whole, what we do in cloud we also do for customer. And I think that's a unique thing that we have at Oracle which is quite fascinating. The operating system we run for our customers, the operating system that's the host part of VMs, is the exact same binary and source code that we make available, just to be clear, the exact same binaries are the ones that you run as a customer on-premises. So if you run Oracle Linux with KVM, you run VMs, you're actually running the exact same stuff as we run underneath our customer's stuff. Nobody else does that, everyone else has a black box. So I think that helps a little bit with transparency as well. >> Yeah, and that homogeneity just creates an environment, you're talking about that sort of security mindset, it's critical, you're not just bolting it on, it's part of the culture. But you started your career, and then of course you were a Linux person when you came to Oracle, but then I think you spent some time in database, back in the day when there were serious database wars going on, before Oracle became the king of database. So now you've got, obviously, this great portfolio, and a lot of really sharp software developers; What should we expect going forward, from Oracle? What should we look for? >> You know, I was talking to some, I was welcoming some interns to the company, for their summer internship yesterday, and one of the things I mentioned to them was that -- so cloud obviously gives us a lot of opportunities, but there's a number of things. One is, we have such a breadth of applications and software and hardware together. We have the servers, we have the storage, we have the operating systems, we have the database layer and so forth, and we have the cloud side, and one of the great opportunities, and I think we've shown a lot of this happening with the ability to create something like Autonomous Database, is to combine all these things. Right, we have such a broad portfolio of really cool technology that by itself is okay, but if you combine the things it really becomes awesome. You cannot create autonomous database without having autonomous learning. You cannot create those two and make them really safe without also controlling the firmware on the hardware and so forth. So by being able to combine all these layers, and by having a really great relationship across the teams within the company, that opens up a lot of opportunities to do stuff really quickly. And having the scale for that. I think that has been, for the last few years, a really great thing, but I can see that being one of the advantages that we have going forward. We have Oracle Fusion Applications, which is incredibly popular, and has great girth, and then we have that running on Oracle Cloud, that talks to Oracle Autonomous Database, so we bring all these pieces together. And no other SaaS vendor can do that, because they don't have these other pieces. They have one area, we have all of them. And so that's the exciting part for me, it's not so much about making my own world better, and having Linux be better, and Casewise and so forth, which is important, but that becoming part of the bigger picture. And that's the exciting part. >> Well, Oracle's always invested in RND, we've made that point many, many times. Whether it's database, you know Fusion was a painful but worthy effort, the whole public cloud piece, obviously many acquisitions, but the investments that you've made in open-source as well, Wim, you're a great spokesperson, and a great representative of the open-source community generally, and then Oracle specifically, so thanks very much for coming on theCUBE and sharing with us the state of the penguin, and best of luck. >> You're welcome. Thank you, thanks for having me. >> Alright, and thank you for watching, everybody. This is Dave Vellante for theCUBE. We'll see you next time. (cheerful music).

>> Announcer: From theCUBE Studios in Palo Alto in Boston, connecting with thought leaders all around the world. This is a Cube Conversation. >> Hi everybody, this is Dave Vellante and welcome to this Cube Conversation. Really excited to have Wim Coekaerts and he is the senior vice president of software development at Oracle. Wim, it's great to have you on. And you know what I often say I wish we were face to face but if we were you'd have to cut off my tie 'cause developers and ties just don't go together. >> No, I know, and this is my normal outfit so this is me, wherever I go. Hi again, good to see you. >> Yeah, great to see you. So of course, you know a lot of people are confused about Oracle and open source. They say, Oracle, open source? What is that all about? But I think you misunderstood. People don't first of all realize you as the leader of the software development community inside of Oracle, I mean, you've been involved in Linux since the early '90s but you guys have a lot of committers. You do a lot, I want to talk about that. What is up with Oracle and open source? >> Well, it's a broad question. So you know, a couple of things. One is we have many different areas within the company that are dealing with open source, right? So we have the cloud team doing a lot of stuff around the cloud SDKs and support for different languages like Python and go and of course Java and so forth. So they do a lot around ensuring that the Oracle ecosystem is integrated in the open source tools that customers use, or developers use Terraform, so on and so forth. And then you have the Java team, and so of course Java is open source. And then, the Graal project, GraalVM, which is a polyglot compiler that run Java and Python and JavaScript and so forth together in one VM, do really cool optimizations, that's an open source project. Also on GitHub, there's of course MySQL which is along with Java, they're probably the two most popular and widely used open source projects out there. There's VirtualBox which is of course also a very popular project that's open sources is all the work we do around Linux. And I think one of the things is that when you have so many different areas doing things that are for that area, then as a developer or as a customer, you typically just deal with that group and what you see is, oh, you're talking to the Java developers so you know what's going on around Java. The Java developers might not necessarily say, oh, and we also do MySQL and we do Linux and VirtualBox and so forth. And so you get sort of a rather myopic narrow view of the larger company. When you add all these things up and there would be one big slide that says, "This is Oracle, these are all these open source projects there". And there's multiple ways, right? One is we have projects that we've opened sourced and all the code came from us and we made it publicly available. We are the main distributor and we get contributions back. There are other projects where we contribute to third party in terms of enhancing things, like a separate the cloud team. And then in general, something like Linux where, you know, we're part of an external project and we participate in the development of that project at large. And so there's these three different ways when you count up all the developers that we have that deal with open source on a daily basis and in terms of contributions, in terms of both Pyxis testing and so forth, it's thousands, literally, full time developers. And of course all the projects is on GitHub or similar sites that are very popular. So yeah, I think the misunderstood is probably a lack of knowledge of the breadth of what we do. And our primary goal is to provide services and products to customers. And so the open source part is sort of embedded in the development methodology, but that's not something we sell or market separately. We just work with customers and products and services. And so in some cases it's not well understood. >> Yeah, well, we're talking, of course we're talking about the state of the Penguin. I think it's part of what people understand. I mean, Oracle got into the Linux game, in the '90s, maybe the latter part of the '90s and Oracle of course wants to make Linux, wants to make Oracle its applications and database run better on Linux. But as you're pointing out you're Linux distro, full support, end-to-end, thousands of people in your open source community and the contributions that you make to Linux, many if not most, they go upstream, everybody can benefit from those. But of course you want an Oracle distro that is going to make Oracle stuff run better. That's always kind of been the Oracle way. >> Well, so yes, two things. The one is that, so everything we do is upstream. So we have no Linux patches that are not contributed upstream. There's no proprietary code in Oracle Linux at all. It's all completely open, publicly available. The source code, the change log, all the commits, everything. It's fully open and public, which sometimes is not well understood, but it's completely open. And everything we do in terms of feature development or functionality or bug fixes goes upstream to the Linux kernel mailers. It's actually, it's the only way to be able to manage a Linux distribution and be a Linux vendor is to live in that ecosystem. Otherwise, the cost of maintaining your own forks so to speak is very high and it doesn't really solve problems. Now the functionality we worked on obviously is focused on making Oracle products run better, making Oracle cloud run better and so forth. However, again, what's important to understand though is an Oracle database is a program running on an operating system that does IO, it does networking, it does memory, it deals with memory management, lots of processes. So for the most part, the things we work on to improve that, helps everyone out, right? It helps every other database run better or it helps every other language run better. So none of these changes are specific to Oracle. They're just things that we found doing performance benchmarks and testing and so forth. But we say, "Hey, if Linux did the following, it would make boot up fast." Now boot up has nothing to do with the database. But if our customers run on one terabyte, four terabyte, eight terabyte systems, and so booting up and Linux starting up and cleaning up memory takes a long time. So we want to reduce that from an availability point of view. So here we're now talking about just enterprise, right? And so there's this broad set of things we work on that definitely help us, but they're actually really completely generic and help everyone customer. >> Yeah, that's great, good. So I wanted to kind of get that out of the way and help our audience understand it. So let's get into it a little bit. What are you seeing, what's going on in IT? Pick your observation space and your vision of what you see happening out there? >> Well it's very interesting. There's sort of two worlds, right? There's the cloud world and move to cloud and there's the on-premise world where people run their systems on their own. And one of the things that we've learned is, when you talk about machine learning obviously is something that's very popular these days and automation. And so in order to rely on machine learning well and have algorithms that are very effective, you need lots of data. And so being a cloud vendor and having Linux in our cloud on tens of thousands or hundreds of thousands of servers or more allows us to have a view of how an operating system works across incredibly large scale. So we get lots of data and so for us to figure out which algorithms work well in terms of, how can we do network customizations, how can we discover anomalies on the storage side and deal with it and so forth, we can do that at scale. And what's interesting is how do we then bring that to on-prem? Well, if we can get the data and the learning done the training done in our cloud directly, then when we provide that service also to people running Oracle Linux on-premises, then that will work. The alternative is to have point solutions where you provide something to a customer and he needs to learn something from small amounts of data. That doesn't work so well. So I think having both worlds on-prem and cloud directly allows us to kind of benefit from that. And I think that's important because lots of customers are interested in going to cloud. Many of the enterprises have not yet, you know, they're starting, but there's still a huge on-premises space that's important. And so by being able to get them familiar with how these things work at scale, autonomy is again important, right? Autonomous database is incredibly popular and so forth. That allows us to then say, "Here, try these things out here. "It's a service, we can show you the benefits right away". And then as that improves, we bring that on to a certain extent on-premise as well and then they can have it in both places. And that I think is something, again, that's relatively unique but also very important is that we want to create an... we want to provide services and products that act similarly on-premises as well as the cloud. Because at some point when people move, we want to make that transition seamless. And what you have today for the most part is one world that's on-prem and then the cloud world is completely different and that is a big barrier of moving. And so we want to reduce that. You can run the same operating system local as well as cloud, you can get the same banality and then that helps transition people over much easier. >> Yeah, well, Oracle actually was one of the... I think but Oracle was the first company to actually market same-same, you actually use that term. Others put forth that concept, but Oracle was the first to announce products like cloud to customer that was same-same now it took some time to actually get it perfective and get it to market. But the point is, and we've written about this is that Oracle, because of the ascendancy of cloud flipped and has a cloud first mentality and you just kind of referenced that you just said, "And you can bring that to on-prem". So I wonder if you could talk about that cloud first mentality and the impact on hype? >> So yeah, I think the clouds first part is of course in cloud we work on services more so than products that we deliver and there's a number of things that are happening. So one is we obviously continue to provide products across you can download Oracle Linux, you can download the database in web blog, you can install it on your own, right? You can do to the traditional way of working. Then in a cloud world, what typically happens is, oh, I use a database service. I'm not installing anything. I push a button and I get an IP address and a SQL, and a connect string and connect to a database. And we take care of everything underneath the database. Now, in order to do that, you need to hold infrastructure in place, right? You need lugging agents, you need a backend that captures all that stuff, you need monitoring tools, you need all the automation scripts for bringing this service up and monitor it. And so that takes a lot of time to do, right? And we learned a lot by doing this. And so the cloud first part of the services means that we get to experience this ourselves with direct access to everything. Now taking that service with all of the additional features like autonomy and bringing that to an on-premises world, we have to make sure we can package that so that all these pieces around it go along with it. And that takes a little bit more time, so we can't do everything at the same time. And so what we've done with autonomous database is we created everything in Oracle cloud, you have the whole system running really well. And then we've been able to sort of package that and shrink it into something that can be installed on-premises but then connected into Oracle cloud again. And so that way we can get all the telemetry, all the metrics, and that allows us to scale because part of providing a cloud service that runs on-prem in the customer environment is that we need to be able to remotely manage that, similar to how we manage something that runs in their own cloud, right? Otherwise it doesn't scale. And so that takes a little bit of time, but we've done all that work and now we've got our customer database that that's really in place. >> Yeah, you really want to have that same cloud experience, whether it's on-prem, in the public cloud, hybrid, et cetera. So I want to explore a little bit more. Who is using Oracle Linux and what's the driver for using it? Can you describe maybe some of the types of customers and why they buy? >> Sure, so we started 14 years ago, right? 2006, October 25th, 2006 (giggles). I remember that day very well. Penguin's on stage and a big launch for Linux in San Francisco Moscone Center. So look, the initial driver for Oracle Linux was to ensure that Oracle database customers or Oracle product customers had a good operating system experience, right? And the ability to be able to handle critical issues when that occurs because typically a database runs the company's critical data. The most essential stuff that a company has is typically in a database, in Oracle database. And so when that thing has issues with the operating system, you don't want just to talk to multiple vendors and have finger pointing and having to explain to an operating system vendor how the database works. In the Unix world, we had a glitch relationship with the OS vendors and the hardware vendors. They were the same. And they knew our products really well, and in the Linux world that was very different. The OS vendor basically did not want to understand or learn anything about products living on top. And so, while, to a certain extent, that makes sense. It's an enterprise world where time is of the essence and downtime needs to be limited absolutely. We can't have these arguments and such. And so that was the driver initially for doing Oracle. So it was to ensure there was a Linux distribution really backed by us that we could fix and we could fully support, right? That was completely the original intent. And so the early customer base was database customers. Database and middleware, mostly database. So but that has then evolved quickly, and so, (clears throat) sorry. What happened was, people would say, "Look, have a thousand servers, a hundred run Oracle, "so we'll run Oracle Linux on those hundred "and we run, something else on those other 900." Now after a year or so, they realized that our support was really good. We fixed all these issues and so then they're like, "Why are we having two Linux distributions? "This thing works really well. "It's runs any application, it's fully compatible. "So we'll just go a thousand with Oracle Linux". And so the early days, the first few years was definitely Oracle database as the core driver and then it sort of expanded to the rest of the estate. And over the years (clears throat), we've added lots of features and functionality, like Ksplice and so forth. We have an attractive pricing model for running on servers. And so now lots of our customers have a very small Oracle percentage running and many other things running. So it's really become a all or nothing play in the Linux space and we're well known now, so it's been actually very good. >> You just mentioned Ksplice. I mean, we've been talking about cloud and on-prem and hybrid and let's talk about security because security really is a differentiator but particularly if you're going to start to put stuff into the cloud. Talk about Ksplice specifically, but generally security and your policy there. >> So security first is sort of what you hear us say and do in everything we do, right? The database obviously security on the Linux side, security matters, Ksplice as the technology is there to do critical bug fixing and make sure that we can apply security vulnerability fixes without affecting the customer and not have downtime, right? And if you look at, most of the cases or many of the cases where you have security vulnerabilities and exploits, it tends to be because systems were not patched. Why were they not patched? Well, not that a customer doesn't understand that it's important, but it's a whole train of events that needed to happen. You have to get notified that there's a security issue in your operating system or application. Then, well, an application typically means it's a multi-tiered set up. So if you have to bring your database server down, then you first have to coordinate with the application users to bring the app server down because that talks to the database. So to patch one system, you basically have to bring down all application stacks. You have to negotiate with the DBAs, you have to negotiate with the app admins, you have to negotiate with the user. It takes weeks to do that and find time. Well, during that time you're vulnerable. So the only way really to address security in a scalable way and reducing that window of time is to do it without effecting the customer, right? And so Ksplice is something that... It's a company we acquired in 2009 and have since evolved in terms of capabilities. And so it allows us to patch the Linux kernel without downtime, right? We lock the kernel for a microsecond, so it's literally no downtime. You don't have to bring down applications. The user doesn't see it. There's no hang, there's no delay. And so by doing that, you can run the Linux operating system, Oracle Linux, and you can be fully patched on a system that hasn't rebooted for three years and you don't even know it. And so by doing that type of stuff, it makes customers more secure and it avoids them... It saves them a lot of money in terms of dealing with project management and so forth, but it really keeps them secure. And so we do that for the Linux kernel. We do that for some of the libraries on up that are critical, like OpenSSL and glibc and one example, I can give you two examples. So one example is Heartbleed was this bug in OpenSSL a number of years ago and so everyone had to patch their SSH server. And that meant basically, systems around the world had to reboot, like a whole active reboot across the world. With the Ksplice today if Heartbleed were to happen tomorrow, we would be able to patch this online for all the Oracle Linux customers without any downtime. No reboots, no restarting of applications, everything keeps running. The amount of money saved would be massive, right? And also of course, the headache. Another example is, (clears throat) and this was an Oracle cloud when some of these CPU bugs that happened a few years ago that were rather damaging on the cloud side, right? Where you could basically see memory of potentially of other machines running that the cloud it's incredibly critical. We were basically able to patch our entire cloud in four hours and the customer didn't know, right? 120 million patches or something that we applied within four hours all online without any down time. And so that technology has been really helpful both for us to run our cloud, but the exact same patches and same fixes go to customers on-premises as well. But this comes back to the whole what we do in cloud, we also do for customer, and I think that's a unique thing that we have at Oracle, which is quite fascinating, right? The operating system we run for our customers, the operating system that's the host for the VM is the exact same binary and source code that we make available, just to be clear. The exact same binaries are the ones that you run as a customer on premises. So you run Oracle Linux with KVM, you run VMs, you're actually running the same stuff as we do for our... That we run underneath our customer stuff. Nobody else does that. Everyone else has a black box. So I think that helps a little bit with transparency as well. >> Yeah, and that homogeneity just creates an environment you're talking about sort of the security mindset is critical. You're not just bolting it on, it's part of the culture. Look, you were, you know, started your career, and then of course you were a Linux person when you came to Oracle, but then I think you've spent some time in the database back in the day when there were some serious database wars going on before Oracle, became the king of database. So now you've got obviously this great portfolio and a lot of really sharp software developers. What should we expect going forward from Oracle? What should we look for? >> I was welcoming some interns to the company, (clears throat) for their summer internship yesterday. And one of the things that I, (clears throat) I'm sorry. One of the things I mentioned to them, was that one of the... So cloud obviously gives us a lot of opportunities, but there's a number of things. One is we have such a breadth of applications and software and hardware together, right? We have the servers, we have the storage, we have the operating systems, we have the database layer and so forth, and we have the cloud side. And one of the great opportunities and I think we've shown a lot of this happening with the ability to create something like autonomous database is to combine all these things, right? We have such a broad portfolio of really cool technology that by itself is okay, but if you combine the things, it really becomes awesome, right? You cannot create autonomous database without having autonomous Linux, right? You cannot create those two and make them really safe without also controlling the firmware on the hardware and so forth. So by being able to combine all these layers and by having a really great relationship across the teams within the company, that opens up a lot of opportunities to do stuff really quickly and having the scale for that. I think that has been for the last few years a really great thing but I can see that being one of the advantages that we have going forward, right? We have Oracle Fusion Applications, which is incredibly popular and has great growth. And then we have that running on Oracle cloud that talks to our autonomous database. So we bring all these pieces together and no other SaaS vendor can do that because they don't have these other pieces. They have one area, we have all of them. And so that's the exciting part for me is basic... It's not so much about making my own world better and having Linux be better and Ksplice and so forth, which is important, but that becoming part of the bigger picture. And that's the exciting part. >> Well, Oracle has always invested in R&D. We've made that point many many times, whether it's database, fusion was a painful but worthy (giggles) effort. The whole public cloud piece, obviously many acquisitions but the investments that you've made in open source as well. Wim, you're a great spokesperson and a great representative of the open source community generally, and an Oracle specifically. So thanks very much for coming on theCUBE and sharing with us the state of the Penguin. The best of luck. >> You're welcome. Thank you, thanks for having me. >> All right, and thank you for watching everybody. This is Dave Vellante for theCUBE. We'll see you next time. (soft music)