In this section, we'll cover what is the benefit of NFV to communication service providers, and what were those initial use cases being virtualized? So NFV value to communication service providers comes in different categories. I already briefly touched on some of those. So let's cover them in a little bit more detail. We talk about service agility and faster time to revenue, which is because now when we have software defined, we can be much faster in provisioning those services in those new locations. We can also have the software innovation. So once we have the platforms virtualized, instantiate the software in the right places, it's way faster than what it used to be, shipping and wiring those physical appliances. For CapEX, once we consolidate the hardware configurations onto a very small subset of those, it becomes-- especially because we talk about IT volume servers-- this helps with the CapEX equation compared to what it was to build all those different individual appliances. And for the operational costs, automation obviously helps. And the next slide will cover a little on how. Using uniform physical equipment is obviously easier for maintenance of this type of hardware environments. And it is also possible if the cloud platform is done with the right separations in place that we can use this NFV infrastructure to run production testing and service upgrades on the same platform. And another reason why this thing is better for operational costs is that finding IT skill sets-- so usual administrators of IT environments is much easier than finding on the market very specific experts on the proprietary elements that it's harder to find. So let's focus, in particular, on how does-- or the major optimizations and cost savings are always being applied in operations. How does NFV help transform it? So in the way how the care networks are managed today, we are looking at a root cause analysis. And the idea here is to move from finding the faults and having enough manpower that knows how to troubleshoot the environments to move towards full automation through different steps. So first is automated watch. And watch here means machine watching machine, so without having human eyeballs involved. This can be done on massive scales using the right environments. Then detect some of the anomalies in the system, decides if something needs to be done, follow on this decision so that the action gets enforced, and this way we can automatically mitigate a number of those problems. And the last step is here is to apply artificial intelligence to those systems and to learn what are the combinations of events that need to be avoided, and simply not allow this thing to happen if the administrators of the network are asking for it. So this will be based on having different layers of the stake automatically detect on the critical alerts and also, the last step based on the analytics. So all this is proven so far in a number of leading communications service providers, both big ones and the smaller ones in size. And here is the list of initial use case tractions up to 2017 in different categories. So to provide consumer services driven by a lot of VoLTE buildout and other types of services being provided, because it's technically easy virtualize control plane, a lot of IMS, TAS, and, as we see, it got virtualized that way. In some operator, this is running in production and networked for three plus years already. And on business category of services, there is a lot of focus on SD1 and providing enterprise customer premises equipment and provider edge and different other similar functionality being virtualized so that, for example, on the edge of the network or on customer premises, we don't layer on top of each other physical appliances. They all run consolidated in some small virtualized platform, which is then energy and cost optimized if its on premise. And then, for wholesale services, there was quite a lot of focus on IoT and MVNO use cases, where the new build out of EPC was done in virtualized way. We also have now, in the meantime, some of the comm service providers virtualizing full core networking, including EPC, for the existing voice of normal customers. And then those three categories were customer facing services. A lot of internal functions in the carrier networks are being virtualized. And here is just an example list of those. Simply decisions for all those categories are being made. If there's new services built more or less from scratch, of course integrated into a lot previous systems, especially customer facing ones, into all this [INAUDIBLE] and OSS, and so on, and for billing. A number of service providers are making these decisions when there is a end of life and end of service. And then they have to face, are they going to refresh this in physical or virtualize? So mostly making decisions to go virtualized because that's the right one. Going on the path to virtualized, automated, and eventually having the networks consumable. So those were the initial use cases up to 2017. And then we have the current ones. And its what is coming next is, for content services, a lot of caching, virtual [INAUDIBLE],, DVR, CloudDVR is being virtualized, already running in production in some networks. In Broadband services, BNG, BRAS, and CCAP is being virtualized also. We do have examples of that in production already. And then, for the mobile services, a lot of different approaches how to transform the radio access network. And what is a huge focus these days, applied across all those distributed locations is how to get edge computing platform and vertical consumer comps integrated with government services also delivered to the end customers.
