I mentioned the over-the-top or the through the network and we just talked about purpose-built platforms. Again, one of the struggles that has been long recognized inside the communication service provider network is that there's a very long lead time for them to introduce new functionality into their network. One of the additional motivations behind NFV and SDN transformation that takes place is to shorten that, is to go to a software-based model similar to what we would think of as a Facebook, or a Google, or Alibaba, Baidu-type of an application where we can introduce new functionality to the network more quickly. By having platforms that are standard high-volume servers and applications that can go on top of those platforms, we should be able to test out the viability of those platforms and roll them out in a trial basis into the network faster than in the days when we had to go to a vendor and acquire a purpose-built platform because it would take them time to build not only the physical infrastructure but the software enablement that's on top of that and deliver that physical platform into that network where we could go and do our trials and our tests before releasing into the rollout. So, again, transforming because of the traffic rate and then shortening the timespan that it takes to introduce those platforms in here. We're going to get back to 5G very quickly but for the moment, let's think about all those elements as how they come into place. So, one of the other words that you're going to hear is fail-fast. I don't think failures are a good word when we're talking in a communication service provider network scope because failure is not one of the options that we really like to talk about, but what we really mean is experimental quickly. We're going to experiment quickly with new functions that may lead to revenue-generating services, and if those functions are proved not to be viable, then we didn't spend a lot of money building having purchased purpose-built platforms and putting them in a network in order to do that so we can experiment quickly. Then obviously, the big motivation is to reduce that time to revenue because of the pressures that the comm service providers are under based on those through the network functionality. So, 5G. Back into the end. We talked before about how 5G is more than just that radio access. Certainly, it's a critical aspect of it where we see that 5G is going to offer millimeter waves. These things operate in the 25 to 30 gigahertz type of range, very high frequency, and that may be in what we'd call a sub-six gigahertz range or that mid-frequency range on the RF side and those radios are going to have to change all that functionality. But 5G is more than just that radio aspect and the functionality, the heavy math that goes into creating those devices that operate in that space. But really is that transformation of this end-to-end strategy inside the network. So, in here, we switch left to right and right to left. So, if we start at the right with the smart devices themselves. So, whether those are the handsets that we carry or whether those are fixed wireless access point, where we can use the very high capacity of the 5G network to replace that short tail that exists in the comm service provider network to get us greater bandwidth for that last mile, with that last couple of kilometers that are in there to, whether it's a connected home or connected factory or some other type of device, right on into the IoT access points that maybe even some of them will have that very long life of five to 10 years from a battery standpoint. The wireless technology itself, as we said, we've got a couple of options inside that 5G use case. It is going to introduce new spectrum into that functional, will use inside to the core of the network. This is a license competition that comm service providers are going to have to go out and go to auction in many cases in order to acquire those licenses nationally. The government is going to license that out to those comm service providers on the macro network itself. We've also got some things that are starting to be called 5G technology and the Wi-Fi in the unlicensed spectrum space and some of the other unlicensed spectrum spaces. Those are typically very short distance in compared to the macro network. They're really good access point, but as soon as we get on the backside of those, we're going to find that either we need to connect those into something like a 5G access point to move that traffic into the macro network as we get deeper into the core network or through some wired connectivity, whether that's a fiber or without a cable type of connectivity, we still need to deliver that traffic into the core, into that access network where we've seen that aggregations spoke about the highways. With 5G, we're going to see, I mentioned the EPC, we're going to change the way that functions because again, we've got some things that the devices are going to bring both from a traffic and from a utilization standpoint as we transform that into that core of the network, and then we finally get to the point where we can provide services into that core network or transport that information much deeper into the network. So, 5G strategy that we talked about really is an end-to-end strategy that includes not only our architectures of our silicon, but the enhancements for accelerated packet processing capabilities like FPGAs, and then moves right on into the areas of security both from a platform level, from an application level, and from a system level, and then the software enablement that takes place, that allows us to do all of those elements and we had experimentation for that quick development of new applications. So, obviously, we're here from Intel, we're talking about this why is the transformation of this network so interesting to us. Well, we transform businesses. This is a massive business transformation that's going on and transforming these types of elements with technology and enabling ecosystems really is in our DNA. So, whether we look at the capabilities that we implement inside our silicon itself, whether that's a virtual technology, the packet acceleration things like the DPDK, whether we've got the quality of service, the amount of resources outside of the silicon through APIs that tell us about the performance of that, vector-matrix instructions like AVX technology, where security applications can be implemented more effectively using things like QAT technology. Then, into the software world of things like our Mac library, the wireless stack, some of our Turbo capabilities in our archives. So, the entire ecosystem enablement that needs to be built from the ground up just like a strong building has to have a massively strong foundation underneath it, and in order to build that structure from the top, we look at those technologies and build it up from that point and deliver that into the community.