 Good afternoon, ladies and gentlemen. You had a wonderful presentation from Intel on 5G. And I'm very happy now to share the view of an operator on 5G. I'm going to share why you believe that automation and open the key interfaces are very key ingredients for the success of 5G. And then I also share why you would believe that the open communities are going to play a very key role on 5G. I'm delighted to have this ONS summit in Europe, taking place in Europe for the first time. Just to remind you, if I may, that the GSM technology was invented here, was built here 25 years ago. And it was built in Europe by people who were convinced that the old, you know, analogic mobile technology was too fragmented, and that with fragmentation, we would go nowhere, basically. And they jointly and collectively decided to build the global standard for mobile and the very key ingredient for this success was to be collective on the technology. Sometimes it's quite interesting to look at the past, to try to predict the future, or to try to build better the future. Let us try to remind what was the key, you know, what has made the mobile technology a success. Obviously, first, there was very huge customer needs, obviously. Then there was some kind of the business side, some kind of virtual cycle, because there was a perfect matching between the needs and the business developments of suppliers, vendor suppliers of device suppliers of the operators, obviously, and all these needs were perfectly matching. This wasn't the business side. Then on the technology side, there was a good matching between, first, there was some kind of global standard, one global standard which was very key to reach the scale. And from a scale, you have obviously economics, you know. And also, it was decided to stick on open interfaces, very key open interfaces. And from all this virtual cycle, we got the success we all know. I think it's important just to remind what was the kind of successes and key successes factors just to keep an eye on the future. So just to be clear, with 5G, we are going to meet another, it is going to be another story, because the underlying technology in 5G is based on virtualization. And the big trend is pushed by a kind of magic world, which is called, you know, this magic world is fragmented, sorry, magic world is the disaggregation. All the technology beyond 5G is based on disaggregation. And the question is, if this magic world is going to help us to build the magic world, you know, this kind of nirvana that we are all waiting for, which is the full potential of 5G. If we are not careful, we just believe that if we are not very careful, we will not definitely not have the benefit of 5G. Let us just remind a bit what happened on the past and have a look at the past. In the past years, we built progressively the 2G, 3G, and so on by putting more and more services. So starting with this all-switched technology and then pushing for the putting some messaging and then some data. Then some on this 4G, we just had the first, what we called verticals with machine to machine. 5G is a huge gap compared to all what we have done and to know. It's a very, very huge gap. So we believe that 5G is not only a matter of spectrum efficiency. It's not only a matter, a question of a new radio. 5G is definitely a very, very, is going to have a very huge gap compared to all what we have in terms of complexity. 5G is going to introduce a new order of magnitude of complexity for the functions we are going to add for the technology complexity and also for all the services and new requirements we are putting on 5G. This is fundamentally why we believe that we will need automation. Why? Just because we are going to have at the same time on the same network a full kind of new services. Obviously all the old services, but also we have machine to machine. We have all kinds of verticals with ultra-reliable or low latency requirements. And we are going to handle that and to manage that in the same network, in the same like a full set of independent or almost independent networks. And this is a kind of extraordinary complexity we'll have to deal with. And this is the full premise of 5G. So let me explain a bit how is that going to happen. This is not going to be for tomorrow. And it's not going to be in one day. Over, I would say that the first phase of 5G is going to be in the continuity of what we know in 4G. So first phase of 5G will be obviously capacity and will be more is going to provide more throughputs, more speed for our customers. This is, I would say, a kind of 4 plus 1 technology. And this first phase of 5G is already fully standardized as 3GPP. It is going to be implemented already in the first equipment. We are at Orangra fully committed to deploy this first phase of 5G. We need this 5G just because we have capacity issues and we need to deploy it wherever we are in Europe because we have capacity issues. But this is a simple, I would say, simple phase of 5G. Just to let you know that we are currently driving field trials in Europe, in our eight countries in Europe. And we will launch commercially this 5G in 2020 anyhow in Europe. Then the big issue is more the second phase of 5G, which is the true 5G with all core 5G. True 5G means that the core network is 5G. It means that it's based on virtualized technology. It is based on this real disruption, which is virtualization. Let me talk about the kind of revolution we are facing here because the journey to the true 5G is going to be long. We are preparing ourselves already. It is a long journey. And to prepare ourselves for the full 5G, the phase 2, 5G, based on virtualized architecture, we already implement virtualized functions on our networks. So we have already some few lessons that we can learn from these first implementations. And the few lessons we have is that we should definitely deal with the risk we're facing. We are facing some risk on this virtualization. And if we do not deal well with this risk, we are going to be out there in big trouble. And all the industry is going to be in big trouble. What we believe is that we definitely have to be very cautious when we talk about the virtualization. 5G is going to have its full potential only if and when the full core will be based on virtualized elements. So now the question we have is, this desegregation can bring a lot of benefits to the operators. Disaggregation has a lot of potential. And at the end, we are fully, I would say, committed in leveraging on the benefits of the virtualization. If we do not handle good this virtualization and the desegregation, we are going to be in trouble. And all the, I would say, the community is going to be in big trouble. So what is the issue? The issue is that we do observe that today there are too many configurations, too many parameters, too many elements that are proposed by the ecosystem, by the vendors. For example, there are at least the NFVI, there is no one NFVI standard. There is no one NFVI standard, one NFE infrastructure standard. There is no today one standard for the orchestration. And the consequence of that is that all the actors, all the operators do have to make more integration. So there are some extra costs to make to integrate the different elements of the NFVI infrastructure. So the observation is that the desegregation is today leading to a fragmented market. And the fragmentation is bad for the ecosystem, the fragmentation bad for the operators. The fragmentation is very bad also for a lot of actors. So if we do not manage, as a community, to make as we did 25 years ago, to make a global standard, we will not have the benefits of the virtualization. And we will have some kind of the mess that we will all know on the IT system. So we are bringing the IT technologies on the network. And the idea is not to bring the mess of IT technologies on the network. The idea is to benefit from IT technologies and also to have the scale effects of the network. The good of the network is that we always kept to one and stuck to one standard. So this is a joint call for action, which we have to try to stick from one standard. Again, if we fail collectively to deliver one standard and to reduce the number of configuration on the yes, on the NFVI, on the orchestration, I really believe that we'll have a kind of industrial disaster. Industrial disaster is one way. The other way would be that some actors will arrive. Some actors will arrive. There will be some kind of de facto standardization of those elements. And they will try to or they will bring their standards and I do not believe that all the community will benefit from this de facto standardization because they will more close, I would say, the standards that open the standards. So now what is the solution and how to scale and to industrialize and how to have the full benefit of the 5G? This is a big question I have. It's a big question we have. And I believe that, again, we have to work 20 on that. For just reminds that Orange, we do operate on 32 countries. We do operate retail market in 32 countries and in 28 countries and we have B2B business in more than 200 countries. And this makes us very sensitive on standardization because standards help us deploy the networks and help us deploy the services of all the countries we have. If we have no standardization, we have to redo 28 or 200 times the same work. So this makes it very complex. So this is the reason why we are very, very pushy and very much involved in all kind of standardization bodies. So we still believe that the right place to standardize 3G, 3G, 4G architecture and to standardize the 5G interfaces is a 3G PPP. Now, the 3G PPP is only the functions, then the question is how you implement it on the clouds. And this big move to the clouds, again, I think it's the journey for five, 10 years. All the operators worldwide are going to put the network in the cloud. So we have a transformation to the cloud of more or less 900 operators, thinking to be a long journey. And we believe that there are at least two key building blocks on which we should jointly and as a community work to have things more standardized. So the first one, again, is a yes. First one is the NFV infrastructure. We believe that one of the right place to do that is OPNFV. Again, open source communities are one of the good places to do this kind of job. So again, I rely upon those communities and we are heavily investing in OPNFV. I think we are one of the first contributors in OPNFV. This is one building block. The second building block is orchestration. We will not have the full benefit of 5G if we do not have the right orchestration layer. And again, the telcos cannot afford to be fragmented. The telco community cannot afford to have one orchestration which is different from one operator to another operator and to another operator and to reduce the integration work. This is not going to be viable. So I urgently, this is, again, a call for action. We have contact with GSMA. We also have contact, obviously, with all the ONAP community. We do firmly believe that ONAP is the right orchestrator for the future. It is going to help to standardize all kind of API in all sides of the orchestration layer. And we urge the operator community to join their forces in order to have ONAP as the target orchestration. So we are investing in open communities in order to address all kind of software and hardware aspects. This is going to be a long journey. We believe that the open communities are the right place to help us reduce the number of configurations to help us finding the right and appropriate standards and interfaces to have the scale effects, the scale that we are lost with the desegregation. So this is kind of summary of my talk first automation and virtualization of very key enablers for the 5G success. It is going to be a very long journey. The open communities are the right place, and we are willing to continue investing in these communities to have the appropriate enablers. And we definitely need to have a standardized NFVI, a standardized or a more standardized TELCO grade, yes, in order to make sure that we get the full benefit of the virtualization without losing the scale effect we had with classical standardization. Thank you.