In this section, we will be covering what are the hardware and software platforms that need to be put in place in the appropriate locations to run these types of use cases. And what is the architecture how does this infrastructure look like. So from Intel, we prepared and invested into a number of different areas. Starting from the foundation at the bottom of this slide, you can see all the different ingredients that we have, which includes appropriate processors. So Xeon for powering usual volume dual socket servers. Or we also have SCC version of that dual power, for example, in the virtual CPE. Then we have it also based on Core processors and Atom processors if it needs to be more power and cost-optimized platforms with lower capabilities compared to what fully featured Xeon can do. We also have Ethernet controllers on appropriate adapters for scalable links. Then we have, based on Altera acquisition, we have FPGA typically coming on PCA adapters to accelerate compute, networking and storage algorithms. 3D XPoint is technology based on which we are building data center SSDs like Intel Optane SSD for data centers. And now also the persistent memory. For similar usage in the storage hierarchy that's now between the DRAM, we will have the persistent memory. Then comes the faster SSDs. Then come the slower SSDs. And then further, we have fixed function accelerators, like QuickAssist that accelerate number of security and compression algorithms. And then we have other assets also coming from the other ingredients and investment in the platform. So based on that, for different ways of consuming those ingredients, we built special software development case that are then optimized to understand and take advantage of all the hardware underneath. And those are being put together into the right reference architectures, make it easier for the ecosystem to build the use cases and to extract value out of all that. So here you have lists of investments happening in the access area, where, for example, we have FlexRAN reference software or we have the virtual CCAP reference software. For the edge, we have MEC libraries, Network Edge Virtualization SDK for developers that allows for easier building of consolidated use cases [INAUDIBLE] then targeting appropriate other verticals. And simply make development effort easier by using those libraries that are optimized for the underlying hierarchy. So there is other capabilities in terms of other libraries like for different encryption ciphers, visual cloud. And then we are building this into different best known configurations, which is recommended and well tested, certified hardware configurations. Which in here, we have example of combination of Intel architecture, so those underlying ingredients with appropriate host operating system, with Hypervisor. And sometimes other software libraries related to that, which allows for easier adoption and consumption of this type of platforms. An example use case that is being worked and also heavily invested is network slicing. Which will allow providing different networking capabilities to different use cases, depending on, some will need super reliable networking on a low latency. And some will require it will be perfectly fine to run use cases on higher latencies and maybe the networking doesn't have to be super reliable. But then from the same physical network, we can slice it and appropriately out of same network expose different capabilities for different customer usages. An example of how this type of ingredients and platforms can be applied to wireless edge computing. So we have specific reference architecture that is based on subset of those ingredients, in hardware. So for example, Xeon and Core processors. Optane memory and Movidius technology coming in platforms with our server OEMs and ODMs that have the multi-access edge compute libraries on top of it. Have the other SDKs like Network Edge Virtualization SDK defined in proper hardware configurations make it easier to build networking functionalities out of that. Then all the interfaces are exposed towards the orchestration layers. And this type of reference architectures then we are working on our own, defining, documenting, and working on adoption with the ecosystem. Some terminology related to that, so Edge Cloud and MEC covers implementations on-premise and off-premise. Various types of connectivity focuses on solving end user, and in particular, enterprise vertical challenges combined with communication service provider value of the edge locations and can include other coms workloads as appropriate. And there is a big standardization effort between ETSI called Multi-Access edge computing. Initially, it was focusing on mobile use cases. Later, all other connectivity was added, along with compute and storage that is required for that. [MUSIC]
