 Hello, my name is Asit Chakraborty. I'm from Huawei Future Network Division in Santa Clara, USA. And here at ITU we have a demo of ICN as a mobile data distribution for ICN slice as a mobile data distribution mechanism. And so what we talk about is what is ICN and how ICN is relevant in 5G, particularly for slices and what kind of applications can benefit and that's what our demo concentrates on. So ICN stands for information centric network. So basically there are many flavors of ICN. CCN and DNR are one of those and they are flavors which use names as a primary mechanism for doing the networking. Instead of IP address we have names of contents and we have entities expressing interest and fetching data by the names. And so that's a very revolutionizing concept compared to the IP address based networking. And it also enables the network to cache data by names. This name based addressing or the name based access of chunks allows the network to cache data. And it also allows many interesting features like consumer mobility is automatically available in the ICN network. And there are other things we talk about like multicasting or the achievement of multicast is automatic also and it's dynamic. So those are the very brief features that we get when we have ICN implemented. So when we implement ICN in the 5G network architecture so many benefits are available. One is the main thing is cellular and fixed access convergence and we can get rid of tunnels and also it's very technology neutral. So any kind of underlying technology can be used and mobility as we will discuss. I mean it can benefit a lot consumer mobility and even what we show here is producer mobility. So these are big benefits and the main theme of this demo is ICN as a network slice. So today I mean deploying ICN is a big challenge and because of the hardware and many other restrictions. So what we try to use in 5G is the 5G softwareization platforms as well as the support for slicing. So what we show is a deployment of ICN in the slice which allows us to get all the benefits of ICN without a lot of interaction to the rest of the network. So you can have IP and ICN slices working parallel side by side and get all the benefits. And what we have implemented here we show a little bit of how the slices are realized. So at the bottom you have the programmable substrate and then on top of that we build up the connectivity slice and ICN slice and on top of that we have the services and the orchestration layers. So we use our VCIR platform for that. So those are basically white boxes which allows the software virtualizations. So ICN itself is or ICN or the forwarding itself is virtualized and then on top of that we have the services and we have containerized all of these components and they are fully softwareized and available in this platforms. And you can have multiple slices and we also support multiple access technologies like LTYFI and Ethernet and we show applications of AV conferencing. And here what we try to show is I mean once we softwareize this ICN under service platforms in ICN we have this interesting concepts called pending interest table and forwarding base and the cache. So when we softwareize all these things can be per service. I mean basically we are having a service specific instance of ICN which allows us to isolate services very well and that's one major benefit of the slicing that we are trying to show here. And briefly these are our demo features like realizing ICN as a slice then how ICN helps in access agnostic mobility and fixed mobile convergence and what kind of applications like as we can support on this. And this is an overall demo setup we have. So we have Docker for container virtualization and we have SDN controller Onos we use for provisioning some of the ICN setups and the service related setups. And on the bottom we have our white boxes which implement the ICN edge routers which implement ICN forwarding as well as the service logic in the network. So in this demo of ICN slices we employ softwareization platform which uses Onos controller for provisioning ICN features like the ICN feed and also service specific setups which we employ in this network. So here you can see this is the Onos view of the three routers that we have in this setup and then this is our orchestrator view. So what we have provisioned here you can see there are the three routers and there are this is the network slice in those routers. So there is a forwarder instance running on each of the routers and there is a service access point and a hypervisor running on each of these routers. So service access point basically helps the user entities connect to the network and hypervisors help the software components to talk to the Onos controller. And so this is the network services we also show the application slice view. So the application here is the audio video conferencing and which uses certain facilities that are provided in the network. So those facilities are again provided with these containers on the network. So we have these concepts of proxies and controllers which are which has been provisioned already and they are on the network as containers. So this is the basic view of this of our setup. Now this is the actual setup as you can see there are three small three of the small hosts are providing the ICN edge routers. I mean again ICN as well as the services and we use one for controlling which hosts all the controllers like the Onos and Docker etc. And we have multiple UVs. So these UVs are connecting to the network to avail of the services being provisioned. So here we show the basic working of the application and so this is audio video conferencing. So each of these participants can be a producer and consumer of data. Right now this application instance is producing data and these two other application instances consumer data as you can see it's live data. So what we are trying to show here is more of the aggregation or the multicasting capabilities of ICN in this application not so much of caching. So whatever data is producing the other user entities are expressing interest to phase that data and that is going through the ICN instances of our network setup and the services instance is helping in synchronizing the application on all the user entities. So this is how the basic application works and this is using right now the Wi-Fi interface. So what we will next show is how it can use LTE or Ethernet also. So this particular instance, UV instance will move from a Wi-Fi connection to an LTE connection. So we have the LTE, OAI based LTE setup here. So the UV is connected to inode B and that inode B in turn connects to the network setup that we talked about. So when we do the handover, so here you can see the traffic move to the LTE interface and the other UVs did not face a very long interaction. So as you see when the other UVs started transmission by pressing the start button this shows the full featured audio video nature of the application. So each one is a producer as well as consumer and we can support many more clients like this. And again this shows the multicasting capability of ICN where the traffic in the backbone, I mean traffic in the network is minimized because of the dynamic automatic multicast nature of name based access. So here we have the visas we call the virtual service edge routers implemented in white boxes and which holds the softwareized containers which implement our ICN forwarder as well as the ICN based services in the network. And each UV produces and consumes a camera feed. I mean the feed produced by him is consumed by the other UVs which have joined the same audio video conference room.