 Okay, let's start. Just a few words, thank you from Nuage Network to invite us to explain a little bit what we've done with them on our infrastructure. I'm here with Fatou Mata, who is our platform owner on network and security, and Maxime, who is the network engineer, the real expert on Nuage. Okay, so the agenda, I will go very quickly on Numergy, because I'm not sure that everybody knows Numergy. Expect all Numergy guys over there, and Nuage people over there. We still have all the guests in the room. So, Numergy had been built two years ago with three investors, SFR, BUL, and KSDIPO, which is the French government. We are now two years after a little bit more than 100 people in the company. We have four data centers. We have one network operations center, one SOC, one lab, and also two POP, which allow us to connect on the internet. I put a lot of different information on the side, but basically we are using different kind of hardware. The main one for the compute is HP, and on the network is Alcatel Lucent with Nuage on top of it. We are providing a bunch of different services. The main one is security, which includes a lot of different kind of small services inside, but we have a SOC taking care for the security aspect of our customer, 24 hours. We are also providing some SLA 49s. In terms of business, we are working only on indirect mode, so we are working with partner, so they are using our own infrastructure to build their solution to sell to end customer. Today we have roughly a little bit less than 100 partners using our infrastructure, and we have somewhere a thousand end customers working or using our VM or network infrastructure. A small slide just to explain and to show you some of our partner and the kind of service that they are providing on our platform. I will not go in detail in it. Numergy is definitely based in France, and we want to stay in France. We want to localize our data in France. That's our main goal, but France is not enough, so we already started to work on a pan-European alliance. So what we do is we look after different public cloud providers in all the country, and we start discussing with them to integrate their offer to us and having one unique European proposition for our customers. We will talk a bit more technically on it later. Just one thing, because building everything that we've done in less than two years, this wasn't possible at all with any real culture implemented in the company, so we are fully agile. We are trying to be more DevOps than everybody else. So that's our main culture inside the company, and that's why we succeed to build what we don't. And now, I will leave the stage to Maxim, who will... It's Fatou? Can you hear me? We use OpenStack as a public cloud provider. We use OpenStack to orchestrate our cloud because of its modular architecture. Well, you know, yesterday the legal example given at the keynotes, I think illustrates it very well. So we can have modular architecture, multi-tenancy, and also automation. And we use it with Neutron because also of multi-tenancy, and it allows us to also have multi-networks per tenant. So OpenStack with Neutron, and we also have an approach with SDN solutions because a software-defined network definitely allows us to have network programmability and dynamic configuration. So it's easy for us when we have a software-defined network solution to automate our provisioning and also for scalability and to manage the overlay and have flexibility and also have a perspective to one oriented to bring all the services such as MPLS for our customers and to extend services on MPLS on our network. So knowing that, I will let Maxim tell you a bit about Neage Network. Hello everybody. So based on these requirements, we decide to choose Neage Networks because you reply all stuff. For example, we can with Neage manage several orchestrators like OpenStack, of course, that we use, but we can in addition use VCloud Director and CloudStack. And Neage Networks can manage several improviser tips like KVM, VMWare, Xen, and coming soon Microsoft HyperV. Neage in addition have many other key features like it can be installed on every hardware so you can use your existing material to install Neage and you can install it very easily. You just have two nodes at minimum to have a lab or for testing purpose. But if you want to have a production platform with high availability, there is only six nodes to install. In addition, it's really easy to connect with OpenStack. OpenStack comes with a core plugin for Neutron and you change it, you replace it by the plugin from Neage. And that's it for the controller node and you install another little software on each hypervisor, the VRS, the virtual routing and switching module. Neage can also provide a new extension of API. You can use Service Chaining. It's the way to forward traffic based on the source IP address or the destination IP address to another VM. You can also configure ACLs by API and you can also configure MPL-S2NL to connect your public router to the Neage infrastructure directly. Now I just will explain how it works when you create a VM on OpenStack. OpenStack, of course, creates the VM and communicates to the compute node to start the VM. Then the module, the VRS, install on each compute node detects the instantiation, the start of the VM and requests the all information you need to build the network for this VM. So the VSD, the Virtual Service Directory, replies through the VSC, the Virtual Software Service Controller to the VRS and the VRS builds everything, all the flow needed by the VM. Actually the flow between the two VM and two hypervisor are supported by VXLAN. So there is a direct communication, a direct tunnel on all hypervisor and with that we are able to separate the hypervisor across multiple data centers. And for the customer it's like the tenants don't see that they are in several data centers so we can have a DRP plan or something like that. So this OpenStack, this infrastructure of OpenStack and Neage Networks enable Neage to be here today and have a production platform since the 1st of July. It's very exceptional because we just started the decision to use Neage at the last year. So we un-start everything in just a month. I think it's very important to say that. Just in the last of April we were with Avana and 2.0 Neage Network versions so since that we have upgraded so without service interruptions and that was easy to migrate, let's say. So I will now explain the numergy design in detail, not exactly but more. So why? We choose of course Neage because it offers us the SDN function but for the load balancing and VPN Neage is not providing this feature and because numergy needs it really we decide to implement a service VM. So I will explain now how it works basically. So thanks to Neage we are able to use service chaining so we create 3 rules just here and thanks to this rule we can force the traffic constrain the traffic to go where we want. So for example if I want to go from one VM to another VM on the same subnet we allow this traffic directly. The traffic don't pass through the SVM but if we want a VM talk to another VM on another subnet we constrain the traffic to go through the SVM to be firewalled directly from OpenStack. So we use with this architecture we can use the direct API offered by OpenStack for the firewall as a service. And we program the SVM by an agent and numergy develop ourselves this agent to be able to configure the SVM. This agent is directly connected to the Neage plugin and Neutron. Because we are a public provider we have several many customers and we have to a big challenge. The challenge is to use efficiently the IPv4 because we don't have much IP. So we have to optimize the way we allocate this IP for the customer. If you want I can be more detailed for this but not now. I can reply your answer how we do it at the end of the session. And the last slide is just to show you the design, the old design of numergy. As you can see it's quite simple. We just have all OpenStack components. We just install in the middle the new edge component and that's all. Finally we just have a basic installation of every component and we are able to offer to our customer a full service of OpenStack. To understand also the service is completely transparent from the customer point of view. Inside the horizon you don't see anything about it. The only thing is that we offer additional services that we are refining with LD and we are providing through the ESM. But it's on the horizon, it's managed by API from the top so it's fully transparent. Yes, thank you. So with this design we can have an upgrade configuration. So thank you Maxime. With the architecture we build we can address every cloud just to offer our customer what they need it. When you talk about disaster recovery plan so you need to store your data and for security reasons if you have some outage you just can restore your data and that's what we want to offer to our customer. And also talking about cloud is about on-demand services so you can have a peak of workflow and outscaling can bring you that flexibility to manage your resources and your usage and also storage and archiving your data. And that's the way we see the opportunity with hybrid cloud is to just change the way you invest and focus your investment on OPEX versus on KPEX. So that's why with all this architecture we build so with Neurage Network and our SVM we are able to address these services private cloud and hybrid cloud. So with security we are building in our SVM and also with our security operation center. So next steps will be very focused on the cloud alliance team so to build an infrastructure that will be transparent and where we could provide and where our partners can provide numergy services with transparency and all the contractual aspects related to it. And also we are looking forward to extend our one with our SVM solution so being able with Neurage to have a global view of your data center hardware I mean routers, your PE so that's what they call data center interconnect and to have a global view from the VSD graphical interfaces and manage the provisioning of your router into that center. And also needs of IPv6 in our infrastructure so we'll go furthermore with all these aspects I think we are done. So don't hesitate if you have questions. We want it to be very interactive and just if you go out you know what about what is numergy, what's numergy experience on Neurage Network and so don't hesitate. Today is basically the entire connection. So that's the first thing. So that's what we focus here. We focus on how we manage to be able to give the connectivity whatever the different kind of connectivity you want and what we want to do then what we've done is basically to automatize everything. Either the customer wants to be connected with a direct private link or through a VPN through internet or a VPN or whatever and everything is transparent. But I'm not sure. The provisioning is very powerful. We have provisioned them outside and then you can make an interconnect. Our world just provides infrastructure as a service. So after it's one of our partner, they have the software or any kind of solution to be able to do this aggregation of cloud and cloud kind of software on the top to manage all the cloud. Examples, yeah. I suppose we'll upgrade very soon. Yep. You mentioned that you have a security operation center. Yes. What different tools does your security center need? Huh, he's not here. Jeremy is here. Probably one of the guys that security can answer. They have a bunch of different tools. What tools do they need? I'll give you my mic. Yeah. So any tools like password to analyze logs needs to like... Every tool you can get is good to analyze. So if you ask what we have, we have like big V's which analyze the connection from administrator. It tries to identify when there is a connection that doesn't sound legit. When it's like a string connection at midnight or something like this. They also analyze logs with a browser, RSI, and vision. It depends on the customer. If the customer wants us to look into, we do it. But for first, we look globally, because we don't want to be too, you know, in terms of the data of the customer. Every event running on our network, on our infrastructure, inside or outside, but it's on our side basically. So we are thinking, starting to launch, something a little bit more dedicated for the customer. So doing exactly the same thing, using the same tool, but for the customer. That means that we need to have a little bit more information about what's going on on the tenant of the customer. So it's a customer choice at the end. Only one. Only one customer. Yes. Today. One question. Clear? If you want more information, just comment below. Both before, yes. Both before. We have a lot of information over there. Okay, thank you. If no more questions, you are free. Thank you, everybody. Thank you very much.