 Welcome one. Thank you very much, Steve. Yeah, Game of Thrones. Anybody knows? Game of Thrones? Yeah? Not everybody, huh? Because if you don't, this will be a utterly confusing speech, I'm afraid, in the forthcoming 40 minutes. But let me check, you know, in terms of your knowledge. So precisely how much does Jon Snow know? Nothing, exactly. Jon Snow knows nothing. If you don't know what I'm talking about, you know, you have to check Game of Thrones. But let me, trust me on this, Jon Snow knows nothing. So do we have my presentation actually loaded? Or, I mean, you know, in terms of the slides? Oh, yeah. Yeah, no, that would not be. So you spoiled completely my build-up. Yeah. So that's good, that's good. So Tuesday morning, right? This is Jon Snow. He knows nothing. Now, imagine that he would have to do enterprise architecture, knowing nothing, because he wouldn't be able to predict anything that business users would need, even a few months from now, even a few weeks from now. Imagine he would have been used to creating data warehouses. He does look like a data warehouse designer anyway. So he would have created data warehouses, but business users would have entirely different needs. There would be sudden opportunities for big database, you know, big data-driven analytics that nobody would be aware of. They want to create new mobile solutions, want to have it in a few weeks, didn't know it right now. And then, being an enterprise architect, he needs to architect for the unknown, right, in an era in which we know nothing. And this could, of course, turn into a bloody mess, like it does in Game of Thrones quite frequently as well. So what I will introduce to you today is a speech that certain parts of it might be well known to some of you. I'll tell you a little bit more about the Open Business Data Lake specification that's currently going on within the Open Group. And the nice thing about this specification is that it is an example of Jon Snow-style architecture. We know nothing. We don't know what we are supposed to create from our data landscape, and we need to architect in an entirely different way in order to accommodate this era of knowing nothing much up front. And that's so-so. I will not only focus a little bit on the Open Business Data Lake. By the way, my colleague Olivier Flébus also with Capgemini this afternoon will have a much more detailed speech on the actual conceptual specification of the Open Business Data Lake as we're currently seeing it. So, of course, I would warmly recommend it to you to have a look at. But I'll put a little bit more in this broader perspective of the third platform or platform 3.0 as we call it within the Open Group and hopefully demonstrate to you that we're very near to a place where we need to create quite a different way of enterprise architecture as well. Now, first of all, I always get requests like this. Could I get these pictures as well because they're so completely enlightening. As you can see, I do like radical simplicity. And these pictures are always suitable to kick off a presentation because they make no sense at all, but they are overwhelming anyway, so everybody wants it. Many IT speakers are like, please send it to me. I'll wow the entire audience just by showing this picture. And indeed, we all, of course, realize that nowadays we're in an era as consumers. We're all very much aware of the social network that we're connected to and all the many different ways that we can get access to real-time information. We're always informed. We're always real-time, right? We're always connected. We're connected to other people. We're connected to enterprises as well. And by the way, if you don't like the circle, I also have the honeycomb which might be, you know, if you think it's more suitable, it sort of says the same. So, I collect these types of things and they're all saying the same things, right? We're very much up to date nowadays as consumers. We're real-time connected. We're connected with everything. We're very smart. We expect immediate responses for the enterprises, from the enterprises that we're dealing with. So, that is a little bit what's happening currently. And then, of course, if you need more pictures like this, obviously, being connected means that there's a lot of different wearables and all sorts of other devices that are way beyond the laptop and the PC and even the mobile phone as we know it and enable us to indeed be real-time connected using wearables or even swallables soon, right? So, we'll have a swallable which has an IP address as well and once a Facebook account, you know, and all of these things start to get connected. And then, of course, you get, I like simplicity, as you see. Then we need to, we see a whole new series of technologies that enable enterprises to keep more or less in sync with what these consumers are doing. So, there will be a flood of data and we're all very much aware of this, the role part of this third platform. There's a flood of data. There's a lot of different devices all being connected. We need to store that information. We need to ingest it. We need to analyze it. We need to structure or unstructure it. We need to make it available as insights so that at least we can match the expectations of the consumer as an enterprise, right? And there's a very quickly evolving landscape very much nowadays, again, driven by the open source community, by the way, which I think is very good news. I have it to be quite a lot in big data in the past two years because of the role I have within Capgemini, the CTO of what we call insights and data. And what you see is a very quickly emerging landscape, in this case of, let's say, big, big data technologies that help us to deal with that completely new flow of data that's coming to us and all the needs of producing insights from it on the other side. Very quickly evolving landscape, a lot of open source technologies over there and what might be fashionable one day like Hadoop, you know, the famous yellow elephant of Hadoop, might be just a few months from now already considered, you know, been there, done that, got the t-shirt, that's something new, right? So it's very quickly evolving in terms of the technology opportunities we have and the platforms that we could be working on. And then, of course, if this is not enough, of course, there's also the area of machine intelligence that is quickly coming as well. So we're adding cognitive capabilities here, artificial intelligence, intelligence machines, all of this, there's still a further and quicker evolving landscape of technologies. And then finally, to give you another of these nice pictures, there are evolving ecosystems of startups. Little companies usually that manage to pick up a few of these technologies I've just described to you that manage to link in into some of these consumer movements we're seeing and they find something new, a solution that might be a breakthrough. It might be a threat, for example, to banks. This is a little snapshot of the so-called FinTech ecosystem which are startups in the banking ecosystem. And they all have tiny little things, tiny little solutions that might effectively, you know, might make an end to, for example, banking as we currently know it. And then it's up to the banks, of course, to understand that ecosystem, maybe be part of it or at least thrive on all of the different solutions that are popping up over there. As an architect, we start to realize that we know nothing because this is so quickly evolving and there are so many new opportunities popping up every day. I think many enterprises shouldn't aspire to be one of these startups. They actually should aspire to curate the ecosystem of evolving technologies and startups, so curate it, find the right components in that ecosystem and use it maybe for extremely quick popping up type of opportunities, right? This is not something you plan for for the forthcoming five here. Let's make a plan and let's make an architecture that enables us to, you know, do exactly the right selected projects in this quickly evolving ecosystem. This is a we know nothing type of era in which we have to take very different measures, also in terms of the way we create our enterprise architectures. So to summarize a little bit, and of course, this is one of the reasons why we're talking about the third platform. My hypothesis, one of the things I really want to make explicit today is that we're not just, you know, on an iterative incremental type of journey right now. There is a real leapfrog. There's a real big change currently going on. And it's very easy to explain. If you looked at the first platform that I would like to call it, this was the mainframe. And some of you are just like me old enough to sort of remember these times. So the mainframe was a very central thing, right? It was an IT driven thing. So IT just said, here's the mainframe. Here's the applications. Use it. You know, that's what they would call, that's what they would say to their business users. It was a very simple era, by the way. They just said, hey, here's a superior type of thing. It's a big mainframe. It costs millions. Here is a simple application. You know, we're targeting specific users. Use it. That's it. Then, of course, in the 80s, we got the second platform, which was the PC. So the PC came and client server technologies. And then later on the Internet, all part, I think, of the second platform in which something very spectacular happened, business, got access to IT as well. And they sort of started to get, they sort of started to understand how to do it. They could have Excel or Debase or whatever, still remember, on their desktops and later on on their laptop computers, right? So they could create solutions themselves. And then in terms of architecture, we got something exciting because we got to do with something very new, which was called business IT alignment. So this was very frustrating and also confusing because now we got business people and they had an opinion about how to use IT as well. This was very disturbing to many architects, obviously. So we had to create architectural ways of dealing with that. So I would say that our current ways of requirements management, our current ways of design and analysis, our current ways of developing systems, but also our current ways of enterprise architecture are all a little bit based on this breakthrough with the second platform in which we needed business IT alignment. So we got demand supply, right? And I had to be a neat bridge between these two worlds, right? Needed to be bridged. So we started with stuff like requirements management and all of our enterprise architecture really geared around this central idea of requirements and evolving requirements as well. So I would say that many of the things that we've seen there, the second platform, are still very notable today in most of our methodologies. Not only in our enterprise architecture methodologies, but also in our development approaches. It's all the very nature of the second platform. Hey, there's business. They can do it as well. Here's IT. We need to align. And that's very, very apparent from what we're currently doing. Now we have the third platform. And essentially it means that there's no longer a distinction between business and IT because the next generation of digital business is actually completely IT infused. It's not something different anymore. It's completely the same. And it's everywhere. There's no longer a PC or a desktop or a laptop or whatever. It's not even a mobile phone. It's everything essentially around us that is part of that connected network. And the speed of change that we're currently seeing, driven by the consumers that are always up to date, always smart, always connected, always expect real time responses. They require a business infused with IT that has exactly the same dynamics. And my belief is that it means that we need a completely different set of methodologies in the end as well to deal with that because this is no longer a matter of demand and supply. This is no longer requirements driven. It's no longer business and IT. It's actually one and the same integrated thing. It's much more like a continuous pulse of change and evolution rather than anything else because IT is everywhere around us. It's infused in everything that we do. We probably need, I think, methodologies and architectural approaches that do exactly the same. And my claim would be my hypothesis is that very soon the methodologies and the approaches and the architectures that we created all from the second platform will not do the job for the third platform. And that is why platform 3.0, open platform 3.0 is such a big deal in the end because it's not the form itself. It is the impact on each and every of the other forms, I think, in the end that really will make a breakthrough and also will point us a little bit towards the direction where, for example, Tocav would be going quite soon after, of course, some of the very important things that we're currently doing. Some people would claim that there already might be a fourth platform lurking around the corner. There are some somewhere which might be all driven by cognitive systems and artificial intelligence and autonomous systems and these type of things. I'm going to bother you with that right now. But maybe a few years from now, maybe 10 years from now, we're still over here in this magnificent hall and we would be discussing the fourth platform. And if we still need to discuss anything by then, by the way, but that's a very different question. So that's a little bit the background and I will use it in the rest of my speech to not only illustrate the open business data lake, but also some of the other things that we're currently seeing. It all has to do, of course, with the culture of knowing nothing. I already had this speech ready and then this Game of Thrones thing started this weekend and John Stowe still knows nothing. And by the way, he doesn't do much either, for that matter. If you know Game of Thrones, you know what I'm talking about. If you don't, never mind. You may want to update yourself a little bit on this, but I think we're currently seeing a culture of know, which means that it's actually an innovation role that you may want to try every now and then as well. If you're in a digital innovation workshop, often I do this a lot with my clients. We just put no in front of something. So what's the ideal bank? Well, what would happen if we would have no bank? How would that look like? What would be the ideal car insurance? Well, let's see what no car insurance would look like. What's the ideal physical product? Well, maybe it becomes virtual and there's no product anymore, right? It's a very powerful innovation rule that we apply over and over again. And if you want to look at the future of enterprise architecture, for that matter, what if we put no in front of it? This always works very well with rooms full of architects, by the way. I have this speech which is called No Architecture, which really creates a warm bond usually between the speaker and the room of attendees. I've done the same with requirements management, by the way. There's my famous speech for them, No Requirements, which really helps a lot as well. But yes, I do believe that we're seeing a culture of know. And already you're seeing it, right? Because what is the ideal data center? We all know what the ideal data center looks like, right? It's of course no data center. It goes away. Everything gets virtualized. Everything became abstract and then virtualized. And then when we had everything virtualized, we could simply move it to the clouds and run it somewhere else. And our data centers became sooner or later obsolete. And this is still happening, of course. Some of you may still have your own data center and you can still put on your shoes with your rubber soles and everything, you know, being anti-static and everything. But sooner or later, these data centers will be empty. And we can create nice, I don't know, nice cubicles for startups, for example, to invite them to come over and do their exciting thing on our premises, right? So that's the future of the data center and infrastructure that's currently happening. If we look at applications, I would say that it's exactly the same thing. If we look at applications, already 10 years ago, by the way, salesforce.com had this logo. Because to them, what is the ideal software? Well, what about no software? You just don't install software anymore. You don't configure it. It just is delivered from the cloud. And the only thing you need at that point is an internet browser. And by the way, for that matter, 70 percent, more than 70 percent of all traffic to the salesforce.com kernel goes through APIs, not through their web front-end. So more than 70 percent of all traffic to the salesforce.com kernel already goes through APIs, mobile front-ends, or Internet of Things front-ends, or whatever type of front-ends. On top of salesforce itself. So applications are quickly becoming obsolete as well. And I believe that if we look at no applications architecture, it's often, of course, a matter of just a set of APIs and microservices that we consider as a foundation, as the platform for whatever application we would like to build on top of it. I know a tech system in Scandinavia, for example, which is completely based on microservices. And the idea behind it is an application. That's really something very volatile, you know? What is an application? Define an application. It might change overnight. So some application is really a matter of combining scripting, if you like, a set of APIs that run microservices. And you create a solution for your needs at that point in time. It might be part of a workflow. It might be a sort of a mobile application. It might be even an old-fashioned web browser application, really, or something else. But the thing is, the real important architecture is the architecture of APIs and microservices. And it's not the actual applications. We don't know. They are Jon Snow applications. What if Jon Snow would be a software engineer? You could ask the same question, right? And he probably wouldn't build applications anymore. He would be building microservices and an API catalog on top of it. This is a very different type of architecture because we don't know what solutions we will need. We know what platform we want to create. And then we would say, here's my catalog of APIs. They trigger microservices. Don't you love it? It's completely scalable. Imagine their business user, all the things we could do with this platform. So we're not going to go to the business users and ask them, what are your requirements? We'll build it. Come back 11 months from now and we'll have it. Probably sort of a little bit. You might not be satisfied. That's a bit the way it's currently done. Instead of saying, here's our platform. Here's our catalog of APIs. There are microservices on top of it. Imagine all the powerful things you could do on top of this platform, which is a very different approach in terms of, you know, it's not demand supply. It's one and the same thing. Here's our platform. It's our foundation for continuous business evolution and renewal. Very different approach. Jon Snow's style approach. And that of course means that open standards are very important. And I don't believe we're very deep right now in the business of APIs, but it's currently happening. Also, of course, outside the open group, I think the open API initiative is an interesting one because tiny little companies like Google and IBM and Microsoft, you know, they are involved in this, but also important API platform providers like Apigee are very much involved in this. And it's important because if we realize that applications become no applications, they actually become catalogs of APIs, the better we understand what an API actually is and how you manage versions of it, but also how to describe APIs even in an intrinsic way APIs would describe themselves. If we have that level of standardization, it truly becomes a platform technology, right? It becomes much more easy to do things on top of that. So this is all happening in the world of applications. And then of course, there's a world of data. You all know the yellow elephant. I hope so. Please tell me. Maybe you don't know Game of Thrones, but you know the yellow elephant. No? Hadoop. A lot of big data people tend to see Hadoop as sort of synonym almost for big data. It's not really true, but it was a breakthrough technology, right? Because Hadoop, being an open source technology, enables you to store and access unlimited volumes of data with almost no structure. That's essentially what we're talking about so that you can do all sorts of interesting data things on top of it without imposing structure or filtering or volume at the very beginning of the life cycle. So the elephant has become a synonym almost to many businesses of big data. And everybody wants to be data driven, of course, nowadays because data is the new oil, right? And if we are on top of having our corporate assets, data, if we can create analytics on top of it and create insights from it, we might be beating the startup ecosystem. We might be thriving, thriving literally on data. That's what many businesses want. So everybody wants to have this magnificent elephant in the house. So everybody's buying these elephants, these Hadoop clusters, and they put them somewhere, maybe in a virtual cloud, maybe still in what is left of their data center. And they also hire, of course, a few of the open source geeks that come with it because you need to run that stuff. And they have a set of data scientists. Imagine, by the way, you put data scientists and open source geeks together in one room is a very interesting mix of, let's say, character building here. Game of Thrones is nothing compared to that. You put them all together in one room and see what happens. The chemistry of it is mind boggling. So companies want to have this magnificent elephant. They want big data technology. But then, of course, if we stay in metaphors, we should realize that elephants can have many different ways of instantiating themselves. So for example, if an elephant goes mad, you don't want to be nearby. So if an elephant goes aggressive and turns itself against you, maybe you're doing something wrong with data and it massively turns itself against you. For example, when you're violating privacy rules or you're using the wrong data and then suddenly the elephant goes wild on you and you don't want to be in the path of an elephant that has gone rogue. So suddenly this big data thing can very quickly turn itself against you. And then, of course, there's a so-called white elephant. Some of you may know this. The Maharajas in the past used to give each other white elephants as a present. This was not a friendly present because the problem with a white elephant is it's so rare that you cannot put it to work because it's a gift and it's very rare. So you cannot put a white elephant to work. But it's a very expensive animal that will eat your entire budget every day again for breakfast. So this was a real big fun between the Maharajas. One of them gives a white elephant to the other and then lasts himself to sleep every night because the other Maharaja was so busy feeding the bloody animal and it does nothing. So this is a thing that we are, of course, seeing with big data as well. We got it. It's very expensive. It's prestigious. It might be boardroom initiative. Now what? It doesn't do anything for us because it turns out to be different than we thought because we are in a don't know type of era. And then, of course, there is the elephant in the room. You know that? The elephant in the room. So the elephant in the room actually is, of course, it's a big topic. Somebody needs to address it. Not on my plate. It's still there. So we can ignore it. We can deny it. It's not working. It's the CIO. No, no, it's a chief digital officer. No, no, we have a chief data officer. Wait a minute. This was a CEO thing. Isn't the CEO responsible for this? We have a poor chief marketing officer. Somebody responsible for this. Elephant in the room. Right? Nobody is addressing it. But still, it needs to be done, of course, because we want to be data driven company. And then finally, of course, elephants could also turn out to be a fantasy. And you know, it's a fantasy animal. It doesn't work at all. We thought the dream would work for us. But in the end, it turns out to be a two-dimensional animation. So not exactly what we were looking for. So that is a little bit what we currently see in the world of big data. There is such a lot of promise. Everybody realizes the potential of becoming data driven and thriving on data. And then if we turn our old ways in terms of, the old way might very well be the mainframe way, the first generation, just buy the stuff, get clusters, you know, get a few open source geeks and a few data scientists, and miracles will happen, which is bad, I think. Might be second generation in terms of business users, tell us what you need. Yeah, we know it's entirely new, and so on, but still, you know, specify your requirements. And we will build a platform and an architecture that will enable you to do these things. But they are all like John Snow. They don't know what they could do with it because they need to be able to work with it and understand because this is a new era. It's a breakthrough type of technology. So that's a little bit what brings us to the business data lake specification and the ideas behind it as well. And as you can imagine, what is the ultimate single source of truth is, of course, no single source of truth. And that's a bit confusing to many data people nowadays as well because they've been trying, literally, for decades to get this holy grill, this brilliant idea of we have a single source of truth and we have it under control. So as long as we have that established, everybody can tap into it and use it. But now actually what we're seeing is a no architecture. It's an example of no, we don't know in advance. So how do we do this? First of all, technology nowadays enable us to get data from anywhere. So whether it's a swollable, that's inside us and has an IP address and once it's on Twitter account still, whether it's a sensor, whether they are ERP systems, whether they are heavy machinery, whether they are cars, whether it's on people, whatever. A current ingestion of data sources enable us to essentially tap into anything. So we don't need to filter anything in this. And it also means that we can literally load everything because these Hadoop style technologies essentially enable us to store literally everything very close to its native format. And that's literally what we're saying over here. We don't know upfront what structure our data will need. So we don't have these requirements neatly specified. So why don't we simply store the data? Even if we don't know if we will use it later on and in what ways we will use it, we'll just store everything. And that means that we also forget nothing. We don't need to filter anything anymore. We can deal with the volumes. We can deal with all the different structures or unstructures if you like, no structures. And also there's no need to forget everything because storage is literally approaching nothing in terms of cost, right? And then of course we can create what we call in the business data lake specification distill points. So depending on what we want to do with the data and we get this moment of epiphany later on, we would create distill points on top of it, on top of this platform, and could do all sorts of different things with it. Could still be our, you know, let's say our data warehousing and our reporting and our dashboarding as we used to know it from the good old times when we still knew what type of dashboard we wanted. And we could specify those requirements and we would create a data warehouse to reflect it and make it available to the business. You could still do these type of things, but we also would have created a sandbox where our data scientists and our business analysts could have a go at. And they might be searching for the oil, right? The oil in that data where algorithms might be hiding themselves that might be predictive or even prescriptive type of algorithms that would enable us to be a extremely responsive business or even being ahead of the pace in which consumers are changing, right? So there's all sorts of different ways of using data on top of it, but the thing is, it's a matter of perspective. The moment we know a perspective, we will apply it, we create a distill point, we make it available. We can do this in a very quick way because it's not depending on requirements. We have the data lake, we have a whole set of technologies on top of it, might be plain SQL based technologies, no SQL, all sorts of different, let's say structures we can impose on it, all sorts of different ways to distill the data, put it in some sort of a perspective, which at that point in time becomes our truth because we all know it's all a matter of perspective. If you look from the perspective of the Lannisters, the world looks quite different. If you look from the perspective of the Starks right now, well, at least what's left of them, this is a game of throne thing. So if you don't know, never mind that. Lannisters are bad by the way, but just saying they're all different perspectives. They don't consider themselves bad, right? It's another perspective on the same data lake. And then, of course, you could provide these insights all sorts of different ways where I believe, by the way, that many insights in the end will not something that you write in an email to an executive and then you take action on it, but it might be turned into an API that hides a very smart algorithm underneath. And again, through a microservice, you would implement it in any type of actionable flow that you would have called in the past an application or a solution or whatever. It could be a very flexible workflow that you create with a web browser. It could be something that is even embedded in the device itself. You'll see in the forthcoming years a lot about edge analytics. And these edge analytics enable you to bring an algorithm very near to the actual source over the data is appearing because it might be streaming and you want to apply the algorithm right at the point over the data is coming in because it might be a matter of microseconds and then we get these so-called edge analytics and quite powerful hardware that enable us to bring these insights very near to the actual point of action. All of this will very quickly evolve in the forthcoming years again. This world of edge analytics is relatively unknown to most people. Trust me, it will very soon turn into an entirely new, let's say, perspective on how we want to get access to these insights and algorithms. And the thing is with a architecture or a no architecture if you like in this fashion we're able to cater for all these different needs on a continuous basis. This is a platform that says here's our data lake. Imagine all the fancy things you could do on top of our data lake which is an entirely different approach from tell me what you need. I'll build a data warehouse and 11 months from now we'll have your modified reports available in your dashboard right which is of course currently still the way we often do things. Again, new standards will be necessary. Again, some of them might be being created right now outside the open group which is fine. For example, if we look at the different versions of Hadoop and the Hadoop ecosystem of open source solutions around it, it's a typical open source way of standardization that is needed over there. Much like we've seen many versions of Linux in the end arise and we needed some standardization between the different packages. And here we're seeing something similar. Open data platform initiative for example is a current, is a standardization initiative where it really focuses on the open source let's say versions of one and the same Hadoop ecosystem. But then of course as I already mentioned there's also the open platform 3.0 that I consider and I've said it before a very important thing because in the end it will have impact if we start to understand that this is a third platform. This is not just a stepwise change in what we've always been doing. This is an entirely different style. There is no business IT alignment anymore. It's completely infused in everything that we're doing and I believe it will have impact on literally everything that we are currently exploring in the open group including security but many other areas as well. So you'll see more about this. Of course we all neatly put it together in open group style as well. So Olivier Villebus my colleague this afternoon will have a more in depth speech and show you a little bit more about what's inside the conceptual specification. We're still working towards the architectural blueprint but the specification is currently on its way. So that is a nice way to tame the elephants a little bit I think you know and let it do exactly what we want despite the fact that it's a very large potentially difficult to maneuver type of animal. I still think it can be done with a no architecture type of approach. Of course there's a few other things that we need to take into account as well. I think in the ecosystem I've used the word a lot. I think stuff like crowdsourcing is very important. What we see in the world of big data is currently that we it's very difficult to hire data scientists. I was at the bank in California and they said we're going to hire 70 of the best data scientists and I told them where are you going to get them from from Google or from Facebook. I get 70 of the best data scientists. Yeah right. You call Mark Zuckerberg give me your data scientist and you'll say yes sure. Sure. I'm sure you can pay it better than Facebook can do right. So you see you see a whole crowd source community right now of data scientists that are free agents and you can gamify or finding your right algorithm by saying here's a test set of data who can find the most accurate best prescriptive or predictive algorithm and it will be gamified and there are literally thousands of the best data science in the world that gather around this collaborative platform and a message over here of course to many companies is the magic probably will happen somewhere outside so you need to reach out you need to be connected again if consumers are always connected with everything including the swallows inside us are connected it means that enterprises need to be connected literally to everything around them as well and if we understand this power of the ecosystem it might become one of our architectural key principles to have an ecosystem architecture if you like which I think is a very important addition that we need and also the agility theme so we want to be smarter through data we want to be we want to be connected as I just show through the ecosystem but then of course we want to be more agile as well and and if you think that scrum and these type of things are or DevOps even as we've already seen it yesterday as well keynote of one my colleagues Gunnar was here yesterday talked about DevOps but we're already looking at at the so-called Spotify teams if you haven't heard of them maybe you have maybe you haven't I hope you've heard of Spotify you don't know Game of Thrones and and no Hadoop and no Spotify there's no hope essentially that's you know just just just try to stay here and stay calm but these Spotify teams you may want to look for the Spotify white paper and it tells something about you know the pace of business change is so fast and it's so completely infused with IT there is no separate team anymore so they're not the scrum IT teams over there with a product manager no the business team itself is responsible for a very frequent business change if necessary multiple times a day in the products and services they're offering and it's completely embedded and infused with the IT change that is part of it it's very much a platform play here's the platform here are the different Spotify teams the the the tribes and the spots working together through guilds by the way it's very much a network type of ecosystem style organization and that's the way they'll be working and the way we will be creating architectures needs to enable these type of teams right so so we need to ask ourselves next to how do we get access to all the data how do we how do we stay connected to the ecosystem we also need to ask ourselves how do we get this type of speed if amazon is able to release new versions of its technology 20 000 times every day literally what if business could do that as well it's often the question I ask you know because they are for sure faster right now then business will be able to change itself so so the innovation ecosystem is a crucial one we're very proud of having opened our own applied innovation exchange in san francisco which is really in the middle of a big of course startup ecosystem the idea over there is not to get into with our clients over there and do some innovation work so bit them but expose them as much as we can to the entire let's say startup and scale up ecosystem as we see it around us because often that that is where the curation needs to to happen and and and if we want to create an architecture that enables that way of thinking and we will meet maybe a startup that could change our lives but we don't know because you know we're a bunch of john snows over here so so we get to this innovation exchange and we want to you know we want to find the next breakthrough and because it's disruptive everybody likes to lose to word disruptive and we don't know what it will be people often ask me you're a cto what's the next disruptive technology i would always say how the hell should i know it's disruptive you know so it's if i knew that wouldn't be disruptive so so you know this that's i don't know so so coke for example they're doing exactly the same thing and and it's all a matter of creating a platform that people want to swarm around this is a really important one i think the word coke is after okay the best recognized world word in the world so everybody knows coca-cola essentially so if they say i have a platform over here it has a business data lake it has apis it has a whole set of powerful source program code source frameworks which startup wants to work with us do you want to work with us we're coca-cola we're over here in tel Aviv one of the places where we you know where there's a really vibrant startup ecosystem you want to work with us you can imagine a lot of people would say yeah yeah that's an interesting platform i want to work on that platform i want to work with such a company in order to create new business opportunities and i believe if we're still talking about an it department or let's say our roles enterprise architects we need to be just as compelling as coca-cola would be in terms of hey here's our platform want to work with it yes of course you know as bees swarming around it in terms of yes i want to benefit from it i see so many opportunities in this platform i'm not going to ask him what are your requirements i'm just saying look at our platform imagine all the great things that you could be doing with it and and this is really the essence i believe behind the approach that we're currently advocating i noticed somebody from nasa in the room as well i like open nasa you know and and we all should be a little bit more like open nasa hey here's our data sets here's our set of apis here's a whole set of uh the code repositories that you would be building on and that you could be using to kickstart whatever you want to build on top of it so here's our kimono right we open it up here's our catalog of things imagine everything you could do on top of it imagine whoever you are so so in the ecosystem understanding what are the different players in this ecosystem what are their interests what are their needs where would where do they want to grow understanding that will be one of the most important things we need to master as enterprise architects as well because this will be i believe the future of of our platforms so there's an ecosystem architecture needed to stay in the business data lake there's a whole very vibrant ecosystem in the lake of course so we need to understand a little bit about these players and i believe if we're talking about real innovation in in enterprise architecture maybe what's what's somewhere lurking around the corner for tokens as well i i believe that the notion of ecosystem architecture will be one of the most important things next of course to the speed and the agility the utter agility that i just showed to you you know with with spotify teams and 20 000 changes a day as sort of an example of the speed that we should be expecting over there but of course also making the apis and data sets available in a way that i just showed to you much more from a a data lake perspective rather than a fixed pre-designed data warehouse right because the metaphor of course is quite a powerful one so maybe because it's funny this crop circle you may have seen it before and the funny thing about the crop circle is that it says requirements management right in the middle isn't that interesting maybe maybe in the near future we need to put something entirely different in the middle maybe it would be something like platform or ecosystem and maybe the cycle over here is very much a sequential cycle right and maybe we need to redesign that as well this will be a very interesting one because what would be the ultimate as i said requirements management might be just as well um and all requirements management so that really brings me to the end of Game of Thrones as well again if you don't know what i'm talking about you you may want to check your classics here we have Tyrion looking at the dragon and i think the dragon is about to spit a lot of fire and we need burning platforms right i think the way that we are working on our platforms which will be the pièce de résistance of enterprise architects i think we will uh we will need to create these burning platforms as a very powerful source of continuous innovation and renewal so let's be a little bit like Tyrion as well and and watch a very intriguing future just ahead of us so thank you very much for bearing with me um please take a seat and we'll uh we'll okay you like to make you comfortable yeah yeah apparently so we'll take some some questions from the audience um but one i have to throw in to start with um when we last heard you talk which i think was this time last year um in Madrid i don't know if i'm good to then you uh you said that an analogy is like a bucket full of water that has a hole in it you can only carry it so far it was for the conic serious in terms of language yes so so today we had a lot of analogies uh in there um try to keep the people awake so early in the morning yeah it's difficult enough i think actually more of them have heard of Game of Thrones and familiar with it them raise their hands they were just it was early morning a bit they didn't want to admit it they were watching Game of Thrones yeah so one i had the all the elephants love elephants the one of Dumbo or the fantasy elephant yeah yeah so the question i had is yes he was a fantasy but the great thing about him was he could fly yeah so how do you get the elephant flying well you know uh yeah that's what you get with metaphors indeed uh that's what you get with metaphors no i so so so first of all i believe too often it is a fantasy and one of the areas i haven't touched on today is of course understanding uh let's say the value engineering behind it because because often if we have that magnificent elephant and we're not able to express the value the business value we create from is we're still in you know unable to really move forward so so one of the things that is really missing in in what i just discussed today so there's the ecosystem there there's this no data and no application type of approach there is the ecosystem but i i think uh the the value the value engineering behind it is is a very crucial one as well and and probably we realize that elephants can't fly so we wouldn't envision a an elephant with uh with you know with his ears helping him to fly but but i do believe uh that's that we need to be much more effective in expressing the value articulating the value of something we want to do it's consistent with what we heard yesterday as well through the theme on it for it okay that's all about the value so uh enough from me um asking the questions from the audience is the open group forum director for open platform 3.0 dr chris harding so chris looks like you have a handful there i well i certainly do have some interesting questions here and i would be about game of france as well if you like i can further explain thanks ron for those for those insights that that you gave us but uh we wish to probe further or the the the audience does i'm going to actually combine some of these questions here if they if they relate to similar topics so uh in fact two here one is is it the end of traditional business intelligence and data warehouse uh and the other uh gartner says that clients with hadoop are achieving about 50 percent satisfaction on deriving business value so i guess the common thread there is is the business data lake the one solution to all data management problems or are there if you like diagnostic situations where you would say this is where you will derive more business value and how do you determine um how that business value could be derived you know first of all i i don't think there's any time soon and and to to what we know as as conventional business intelligence and and decision support because all of that is is still of course with many companies something they yet need to establish the way i see it is it's yet another perspective you put on top of of this platform that would enable you among other things among many other things also to create your your let's say your monthly dashboards and and and you're reporting even for compliance purposes for example there there is a very specific um perspective you put on top of it and we've already found with companies for example like unilever like that we've done over here in the uk that that you can speed up that process as well by by implementing such an architectural foundation so so you would be able to produce reports much faster instead of the you know sequential requirements driven approach to i need a change to a report or i need a new report and hey maybe we need to restructure our data warehouse you know and i need to rethink our data ingestion and our integration and our etl and it's you know it's a very linear let's say requirements driven a thing that that's you know stipulates that we need to understand structure up front anyway we're going to deliver the solution and now we would say no we would be very quickly able to tap into that data lake create exactly the data set that we need for example in this case plain sequel in in some sort of data warehouse denormalize scheme make it available and do it if necessary in in hours or days rather than in the months that often with with these companies used to take in the past but this is only the classical thing and because everybody realizes that the new oil might be in hidden algorithms and we need a sandbox to unleash our data scientists on and and find you know the new value in that data as well which is often not in the bi and in the reporting and in all the compliance you know data that we that we present so we're talking about new value and if i actually one of the other questions talks about the importance of velocity of value is that could you perhaps expand on whether velocity of value is a is an important concept and and how the data lake helps to achieve that well velocity of course is one of the famous v's of big data right you have your volume variety velocity and also value and even a few other v's people you know pride themselves in making up yet another v behind big data but velocity of course is a very important one because in many cases you'll have a you'll have a stream of data coming to you and you need to be able to apply your algorithm if necessary in microseconds to that very stream and and and be able to do something for example if a car drives itself you may want to refrain from batch batch processing let me put it this way in terms of your analytics right needs to be done in real time but we still want to collect the data later on because we may want to do some of our data science magic on it and find even better algorithms to to to you know steer the car forward so it's a matter on one hand of microseconds and being able to apply the algorithm right on the spot using for example edge analytics and on the other hand we want to have access to that data maybe two years from now because suddenly we realized through a new hypothesis that we could do something with data that we once captured never realized what what value would be in it and we apply some of our algorithm magic to it and we find something new so you know very different velocity microseconds and maybe two years from now right so that the one on the same platform dealing with that right so velocity is important but it's not the only it's not the only time imperative the long-term value or the available oh yeah absolutely could be something we suddenly see two years from now is absolutely yeah okay so thanks for I think you've you've you've clarified that point very nicely if I can move to the next one sure um so um data pushed into data lakes often lack data models is is data models models yeah that's there's another perhaps relevant to that question when you have all of the data available how do you find the data that you need well it's a it's a little bit often so if you're doing your sandbox style magic it's up to the data scientists because they are data scientists right and they're supposed to be able to find they are they're supposed to be able to articulate hypothesis and but but of course you also see already a lot of automated tools coming up as well that help you to to go through that data lake and find hidden patterns and sometimes there are simple things like like IBM so Watson analytics is a nice example this is a tool that enables you to even do a self-service data science do it yourself data science citizen data science if you like so so you don't need to be a trained data scientist because they're relatively scarce you know and they might all be working for google indians but but still you you could be assisted by tools help you to find hidden patterns in that in that if you like ocean or lake of data that that might be hiding that that crucial algorithm that will make a change later on so so the more people get a feeling for that the better i don't like the idea of leave it up to the data scientists and this person needs to find it i like the idea of democratizing it and and if everybody tool supported probably would be able to find their own way in that data lake find maybe a new hypothesis a breakthrough idea then maybe handed over to the professionals because not everybody can be a data scientist handed over to the professionals that could make sense of it might be very interesting vibrant type of system again in which everybody is a little bit of a data scientist which which i like as an idea so that sounds actually like a big philosophical change instead of going into the process with a predefined model you you put the data into the data lake and the the model or maybe different models come out absolutely of the analysis process the ultimate model is of course no no model yeah no we shouldn't we shouldn't impose a model up front so this is of course one of the biggest changes but it's a big architectural philosophy as well don't impose a structure upfront we don't know so forget a lot of things that you thought you had to know upfront but create a platform instead and be prepared to cater for whatever needs will arise which is i think a very different philosophy and how does that relate to if we go go into the metadata is is is there a parallel thing of rather than metadata being created upfront it is derived as the data is is analyzed or yes could we be looking at although although it helps of course it helps a lot i see a lot of very good examples right now for example microsoft technology that currently does it in which you crowdsource identifying and categorizing and naming your data so somebody would find something interesting in terms of a data set and say hey i think this is what it is about might be an external data set that you suddenly have available and you're like hey that's a very interesting one i think this is what it means others would pick it up in a crowdsource type of environment and would enrich it and would say yeah maybe it's even more like this or maybe we should add this to it and you'll get a very lively again ecosystem network style of enriching the data through through adding metadata and and let's say adding more meaning to it which which again is not something you should outsource to to a single person or a single business unit but make it a little bit everybody's thing to be involved in that as well that that's a real data culture that that i think everybody needs to be aware if it this is the new oil you better enable literally everybody in the organization to to contribute to it okay and again that's an interesting philosophical change of emphasis that the data lake is is bringing into to architecture if i can if i can move on to another one one more problem Chris is all we've got time for it yeah um let me move to one more question then um which you sort of touched on a little before do you think data lakes will be impacted by new privacy regulations requiring us to know and state why we store data yeah absolutely and again by the way perspectives might be very quickly shifting over there i've seen quite a lot of examples in which um you know um on one hand some of the existing privacy rules and regulations of course ask us to apply certain perspectives to the data and again by the way it might be just a matter of perspective because sooner or later we might say hey you know maybe maybe that these rules and regulations need to be adjusted to to mirror what we're currently really seeing in society right it's the same with self-driving cars right now it's considered something that you really maybe shouldn't be doing and 10 years from now you probably have to explain to the regulator why you think you can drive yourself no really 10 years from now why why why do you want to drive why do you think you can drive yourself well i did it in the past you know so and nothing happened yeah well we had a few accidents a few people got killed yeah sure well we don't want that anymore so you're not allowed to drive yourself you know and it's it's really a tough thing later on to your grandchildren you will be able to say i i'm from this era in which we drove ourselves and you're one tough cookie right so so these perspectives will will shift over time and my point again is if you if you design an architect your your your your your data landscape to to reflect what you currently see it might be very inadequate in dealing with including rules and regulations that will be imposed on you pretty soon and we don't know what directions it will go because drones might be currently forbidden and then a few years from now you know they might be all around because they're autonomous and and you know they would be able to avoid each other much better than than humans might be able to do it so so this is a shifting a set of shifting perspectives and again there's no single truth i believe it will it will continuously evolve and that's what we need to architect for to deal with that okay grace thank you uh ron thank you