 Yep. Hello, Leon. All right. OpenStack demo theater. It's pretty exciting. Okay. Ready to go? We start whenever we're ready? Start? Awesome. All right, guys, we're excited to introduce ZeroVM today. This is our startup. We're excited to be here. OpenStack is our place to announce our company, and we're going to go through a brief demo here. My name is Michael. This is our CTO, Konstantin, and we're going to show you some good stuff that we've been working on. So every modern cloud provider that you see here, everything that's done, it's all built on old-fashioned hypervisor technology. It's all built on stuff that predates the days of cloud when people's servers lived inside of their enterprise. So we've taken a look at that and realized a lot of what people are doing from a hypervisor perspective is like using a butter knife for a screwdriver. It kind of works, but it's clear there should be something better. So we're proud today to introduce our product. It's ZeroVM, and it's the screwdriver for the cloud. It's the first hypervisor built directly for the cloud, and we're going to talk about how that works. So before I give you more details about it, let's talk about a use case. So if you're in cloud today, you have your application separated from the data. Okay, let's say that you're going to do an intensive data processing application. A request is going to come in, it's going to go back to your data store, and you're going to pull data back and forth each direction multiple times. This is what we do every day, but clearly that could be more efficient, right? Why are we taking all those extra hops? So imagine you could co-locate some of your code inside of Swift, okay? And that's what we enable people to do as part of our product. So this is just one use case, but you can see how we're thinking differently about how cloud should be put together. Oh, cool, animation. Nice. So what is 0VM at its core? It's a super lightweight hypervisor. It's a 75 kilobyte executable. It provides a single threaded C99 compliant environment, and the security is derived from the Chromium NACL project. We've taken that and repurposed it for use on the server side and add all the stuff that you need in the cloud. Cool, why are we different? What's so special? You can boot up a VM in less than five milliseconds with us. You can isolate your users. You can have one VM per user. No longer are you sharing one Linux server instance with 50 different clients, with 50 different users. One user, one VM. Efficiency, we're only exposing the things that you need. You don't need a TCP IP stack necessarily. You don't need a file system. We're exposing only what you really actually need. Cool, I'm going to turn it over to Constantine. We have some examples of how we've deployed this product already, and here we go. Oops, it stopped running. Yeah, there you go. And Constantine. Hello. I will try to explain the demo. Right now, we are doing login to our own web UI. Here, we're running PHP inside Swift. Our own web UI access is Swift directly. You don't have any type of web server between Swift and web UI application. It's just JavaScript talking directly to Swift. So we are uploading PHP file and PHP executable to Swift. You can see the type of the executable is PHP application. And this is what inside, usual PHP info. We'll create couple of containers and object in Swift to make it possible to execute things. We create application and PHP directory. And we upload the configuration file that it will describe what we need to execute when PHP program is opening, or PHP file. Here's the configuration file. We want to run PHP CGI application and want to do it out of PHP container. This is the object we have here. We uploaded it before. And now, if you open PHP, it just runs it inside the Swift. This is the execution time. It spawned VM, it executed, it brought back the result instantly. Another application is to screenshot from video file. We create screenshot container. We add the executable and the video file. Because we are talking directly to Swift, we need to, with the browser, sometimes we need to say audio instead of video or some other type of content type because we cannot, we don't have any middle layer between the systems. So sometimes we need to make it up for browser incompatibilities. So here we will take screenshot of 21 seconds of the video inside the Swift just by opening the video itself. So we just try to open the video and we get screenshot instantly. Grip. Right now, you cannot search things inside the Swift. If you have a lot of log files or text files inside Swift, you cannot do anything without them, without pulling them out. But right here, we will just search files inside Swift for some specific regex and it will get us the result outside of the Swift itself. So we'll upload some text files. It's 10 500 meg text files. It takes some time. We got five gigs of data, of text data. And we want to search it. Here we upload the configuration file that will launch the search for all these files. We will search for the word gravity using the executable and inside all the TXT files we have in the container. So just by running it, we will get the result. In two seconds, we will just grab five gigs of data inside Swift. We can run other executables like Python, SQLite, any type of executable you can compile for zero VM platform and run it. For example, here's some Python code. We just open and run it and we get the result immediately. Here you can see all the Python imports. It's imported to our platform and executes almost instantly by shipping the Python to the Swift cloud. For example, we want to execute SQL file. So we do select from something. We want to use it open and we just instantly get the result of SQL. We have six minutes left. Image processing. We have here image. This is a JPEG. You see this place is empty. When we open it, we just run a watermark application and it places watermark exactly on the spot we want. Just by opening the image, we can execute stuff and we even can change it on the fly. So for example, we want it in a different place. So we change the gravity and we got image in different place on the bottom. So it's not a pre-compiled image. It's the image that instantly executable, like your data is executable. You can execute code when you fetch the data. Same thing here. It all works on top of vanilla Swift. It's nothing more than Swift itself and WebUI runs in browser. So it's Swift and browser and nothing else. Yeah, we change gravity here also and we pass the watermark to another place. And the final thing is benchmark of Terasort. We will run MapReduce job inside Swift. We have input Terasort files, four gigs of random data and we have the output is empty right now. We'll run it. It takes some time. We just speed it up things a little. And in slightly more than one minute, we did a full Terasort of four gigs of data inside the Swift without any other application. You can see more demos in our booth. So please come and visit. Cool. Thanks for listening. We are integrated into Swift today. Our booth is right over here on the left side of the stage. Please come see us. We'll be happy to talk to you more about the product. We welcome your feedback as well. Our email address is up here on the screen. You can download it today for free. It is open source and available to anyone. We look forward to talking to you about the product. So thank you very much. Have a nice day and a good show everybody.