 Okay, hey good morning everyone First I'd like to I'd like to thank the the Linux foundation events team Angela and crew I've been I've been speaking at Linux Foundation events for For many many years and and you know, it just keeps getting better. You know, so just hands For Angela and her team for an amazing job Thanks, Angela Okay, so every time I speak it's really an honor for me to be I manage a large group of open-source engineers at Intel Linux and systems and open-source engineers and and it's really an honor for me and I'm always thrilled to speak on behalf of those engineers because all I'm really doing is just I'm gonna be talking about some of the coolest projects that that they're working on and I'll talk a little bit about What they're up to these days, so So with that, you know Let me let me start with a little bit on what is What is guiding our vision In a lot of these projects that I'm going to talk about so so everybody here is very familiar with With software defined infrastructure in the data center, you know software defined compute orchestration software defined storage networking Great success and this by the way would not have been a success without the amazing open-source projects I mean literally, I don't know who made the comment about open source made the cloud That's absolutely true. It is it is Kubernetes. It is Linux. It's KVM It's it's all of these projects that made, you know, software defined the software defined Set the data center the dynamic self-service the data center made that possible. So we believe that that the future is software defined everything and By that we mean that things like industrial automation factories cars planes turbines all of these things are at some point will move Slowly to be software defined and there are a lot of advantages to that and we also believe that just like what happened in in the cloud the same thing will happen in all of these software defined segments Segments and usages and that is open source will lead the way and in making this This almost revolution Happened and in compute and in all of these usages Before You will hear me talk or mention clear Linux as I talked to through a lot of projects So a lot of you probably familiar with it I'll I'm just gonna spend a couple of minutes talking about the clear Linux projects just to make sure that you know As I refer to it people just know what I'm talking about. So clear Linux project is a Linux distribution that is developed by Intel and And there is nothing new about Intel developing Operating system. I mean we've done we've created the octo. We've created Zephyr We've created a bunch of other distributions Linux and and and other open source distributions, but what's different about clear Linux is is a project which we started about five years ago is it is really catering to the modern usages modern usages of Cloud of software defined Modern usages in in an IOT and in tether devices in general so going beyond the traditional embedded and and and the traditional Traditional data center. So this is where we Do the work for things like functional safety adding the functional safety capabilities Which I'll talk about a little bit more about how do we utilize the hardware, you know How how we evolved it to a super super fast cadence cadence of a of a Linux distribution in days and and our purpose is is really for For everybody to be able to take from this distribution and Steel with pride, please do just take whatever you want from and we will help you do that and Whether it's features or take it as is, you know, we want you to actually use that so the the key thing The key thing that You know we the key conclusion a key conclusion that we came to Especially over the past couple of years is that for this to be usable by others We need to develop it as a product quality Distribution so that whatever code comes out of it is product quality. Okay, so with that Let me talk about one of the first things that we've been working on for quite some time for quite some time and that is safety critical systems, so So safety critical systems as I said, we believe that a lot of these industries whether it's cars or medical devices or Or industrial Automation will migrate to being software defined Well, that brings in a new problem a new problem that we did not necessarily have in cloud and in data center and that is the safety criticality or the functional safety of an operating system because Because with with this kind of a usage you have you have things like that are truly functionally safe and and by that by the way what it means is that There is some kind of an injury or something bad really really bad will happen in the path of functional Functionally safe and there are things that it's okay. You know if things bad happen So to give you an example if there is a latency in your video. That's kind of okay It's annoying, but it's okay But if there is a latency in your braking system you could die so and so with a software defined system that would use Something similar to this it is it's really important to figure out. Okay, what is all the safety? How do you make something that is open source safety critical? So the industry over the past, you know, 20 years figured out how to do safety critical through microcontroller systems So it would be microcontroller. That's a safety critical then the rest of the system is not microcontroller So right now we're saying that hey This is going to be software defined and you have to be accommodating both So you probably heard me talk about the Intel safety critical projects that we've been that we've been working on probably for a couple of Years now we talked about it a little bit. It's been worked mostly internally and and and this is This the the key thing to understand about the and we started it by the way with the hypervisor part This is the acorn hypervisor that we launched with the Linux foundation few months ago And now we're moving into the OS which is the big piece of you Well to that that will create the the entire environment to be able to run a functionally safe OS so The key aspect of this is that it's not so much about code It's all the components and and all the argumentation That is intended to make somebody create a functionally safe system And that is really really hard work and without this you will not be able to have a fully software defined Whether it's software defined car or software defined anything So so this is a hard problem and this is where we really invite you to start participating in our safety projects on clear So and you will hear us talk a lot about more much more about this now second problem that we've been facing and and all of our you know Intel customers and partners and and quite a few people from the community is When we have these discussions around hey Intel is doing this feature in hardware or there is this feature in hardware forget it and the problem that That we're beginning to see given the complexity of the features is That nobody sees any benefit from those features not really not unless you are a Company and you build an entire vertical solution. That's when you start seeing a benefit So we did this experiment with with what is called an AVX AVX is an instruction set that is That's advanced vector Advanced vector extensions in Intel CPUs So what what these what these instructions do is allows you to do math like really quickly in in very very few instructions so Intuitively you'd think that Something like deep learning where you are doing a lot of math a lot of matrix Multiplications and matrix manipulations you think that oh my god AVX would be like make it scream The reality is so some of the kernel enabling and some of the various components You know work was done for AVX and various components and nothing happened. There was no no improvement So so what we ended up doing is taking a look at the entire stack what's what what you call like a deep learning as a service and Every single component top to bottom and then we went through every single one of them and enable just utilize the hardware For that purpose and you know what happened over seven percent performance improvement This is not a two you know twenty percent performance improvement or fifteen But this is seven over seven acts performance improvement when you utilize and you look at the use case You know top to bottom the same thing we did with With with Apache Spark same results eight acts performance improvement So that got us to thinking and creating you know these use cases that we are trying that is built on a common That everybody agrees to a common infrastructure as a service with you know basically Linux Kubernetes you know KVM so everybody agrees to the basic Basic infrastructure as a service stack and then build use cases on that and we're still you know Much of that work happening internally You know obviously we'll bring that out and and and we'll make it available But this is what really excites us about this and it's not only about use cases when it's performance when you really take into consideration of using the hardware correctly. These are just a Bunch of benchmarks from for on X you know these are public benchmarks You can take a look at them So one of the things that we do in clear Linux and we encourage all everybody who's using Linux to do is to be Aware of what the hardware can do for you. So in this case, you know in clear Linux, you know Just the latest some of the latest benchmarking you have some something like the R language that is you know 4x 4x better some tensor flow 2x better So when you use the hardware in those cases you actually get a real benefit from from these Utilizing the hardware properly and so please go ahead and take a look at clear and what we are doing there and still with pride Now I'm gonna switch topics and talk about application isolation and and isolation in general so Some of it you're aware of I'll just like to be the ad for our engineers and this is around what we've done with With Cata containers, you know the choice that we were faced with or container Partners or container users were faced with is that you had on the one-hand containers great They're very fast. They're very simple to use and on the other extreme You have full virtual machines which very very secure but slow So so you had the choice of speed or security and what we really wanted was speed and security So this is what we developed Intel the clear containers which we eventually you know working with You know Huawei and Microsoft and Google and a whole bunch of us got together and created this merge of a project that is That's on that same concept that is Cata containers that we that we use You know that is containers in the sense that they are, you know, they're compliant with OCI It is it's integrates with all the all the cloud infrastructure and so on and what that got us to think is about virtualization technology So virtualization is very very old as all of us know. It's wonderful It's the basis of much of the data centers today but also it's still it's still and that It hasn't grown it is it's slow. It's It still supports floppy disks and I don't think there's a lot of floppy disks and in data centers so so we we started taking a look at that and A new project emerged a new project emerged that is that really separates the emulation From virtualization and it completely and by the way like code size goes down dramatically very very dramatically complexity goes down dramatically performance, you know significantly much better and This is basically defining a free PC legacy, you know Virtualization infrastructure all KVM and and really moving that forward. So this is just getting started It's the project is on GitHub feel free to it's a GitHub Intel NEMU new EMU. So please do feel free to to take a look at it It's it's probably one of the most exciting projects that That we're seeing today. So and finally on application isolation I wouldn't you know, this talk would be missing given that I spoke at about functional safety without talking about the functionally safe hypervisor because that is what is the underlying foundation of Software-defined Infrastructure outside of the cloud as I talked in factories cars and and and so on so 25,000 lines of code hypervisor that is functionally safe capable with with all the capabilities that is designed specifically For that purpose So much of our work we're bringing to now extending to a lot of other projects so you can hear all about this and You know, there's a bunch of sessions probably one session I do encourage all of you to go to is is I net a session on clear Linux. It's at four o'clock today and and in our booth, you know plenty You know, we're talking, you know about all of these projects acorn clear Linux and kata and and and whatever other things You'd like to talk about we're there to talk to you. So thank you very much and have a good rest of summit