 So our first speaker is Zohar Bevin from Kaltura, so he will introduce us with the 10 years of open source innovations at Kaltura in online streaming. Thank you. So my name is Zohar Bevin, I work at Kaltura, an open source video platform. And ever since 2006 we basically looked at the world of online video and said there needs to be something for video at large scale. What we'll do today is basically kind of go through a 10 year brief history lesson, if you will. But then through that we'll talk about what Kaltura is and what we've done so far and we'll finish with a look into the future and what we're working on and what we would love your help with. So 10 years and under 10 minutes. So interestingly those should be check marks. So, good. Our story begins in 2006 where basically YouTube did sort of transformation in how the world sees video and how people behave online. The world basically shifted from text to video in merely a year. Basically YouTube launched, YouTube reached millions, every schmo is now, you know, I need it with a camera, I can be a VIP. And suddenly everybody cares about video and bam, Google buys them for about $3 billion. So what's interesting is, you know, what we started saying is video is mostly, video online is mostly driven by user generated content and by advertisers. This is where the majority of video online kind of existed. It wasn't just YouTube, but it was mostly YouTube. And that was basically the premise of Kaltura. So there's some numbers here, but some interesting kind of evolution of online video. But then where Kaltura was launched was at the inflection point of video is becoming a standard. And as video is becoming a standard, there's also a standard in technology for video. So the video element is being proposed in 2007. Kaltura.com collaborative video editing is launched. And basically the premise was there's YouTube, there's Wikipedia, but there's nothing that does both. So you can upload video, you can share video, but you can't really collaborate on video. You can't really edit video together. You can work with people on video online. And the premise was there's a huge opportunity here because lots of people are going to use video in the future. And it's going to be the dominating kind of data asset, the way that we communicate online. And we need the capabilities to actually handle this at scale. So 2008 started alongside proposing the video element by Oprah back then. What was interesting is it didn't really catch on. So the first browser came in, they said, here's the video suggestion. So how we see video online, this is the suggestion for HTML5 video. Nobody really cared. Nobody really went ahead and implemented it. There wasn't really a standard around it. And there was not enough driving force. So we went ahead and actually launched together with Mozilla and the Creative Commons and Yale and a bunch of others, an organization called Open Video Alliance. The Open Video Alliance had a few conferences, the Open Video Conference. And the goal was to basically bring everybody together in one room to talk about the future of video to create open standards around video. Alongside, we grew as a platform. We grew as a kind of a place where people collaborate and place where people create video together. We did all sorts of interesting projects together with Mozilla, together with a lot of different creators online that basically went ahead and created documentaries where lots of people around the world could actually upload video and collaborate online. We've actually built this video editor online and so on. And the year ends with 2009, we started seeing video being adopted in other places, not just online in the public sense, not just the UGC slash advertising environment, but also in education. So the evolution is now a revolution in certain industries and in the world of education, we're actually starting to see revolution of education just because of video. Remote learning is now a thing. Flip classroom is now important. Professors are actually starting to take their classes online. And MOOCs are actually starting to be born. Another interesting thing is the outcome of those collaborations with Mozilla and companies like that. And this is actually clickable. You can, after that, if you want to kind of go through the presentation and actually see the videos that you can click on of people. This is Mark from the Mozilla Foundation talking about the Open Video Alliance and why it's important. Anyway, the outcome was adoption begins. Browsers are actually no care and they're starting to implement video natively. Flash is now being treated as a fullback, not as the main tool. So 2009 continues starting 2010. What we're starting to see is video is getting more important. It's getting more embedded. It's getting more native in different sorts of applications. And the more video you create, the more metadata you need to produce out of it, the more complicated the workflows become. It's not just about uploading video from a user device. It's about uploading video from mobile devices. It's about ingesting and scheduling a video from multiple different places, multiple different solutions. Like here we have sort of a lecture-capture box that the guys at Fosden built. So there's a myriad of sources of how we ingest content. Fast forward for today, we actually have lots of different video sources including dash cam that policemen wears, for example. Alongside that, there's also lots of sources for metadata. So how do you produce metadata from that video but also the metadata that correlates to that video? Where was the video taken? Who took the video? In what context the video was taken? Where are IoT sensors around that video that correlate to what that video is showing? And so on and so forth. So now it becomes sort of a challenge of managing lots and lots of data. And it's a complicated data because oftentimes you can't really extract the information easily. And it becomes a big data problem. It becomes a complex problem of how you deal with multiple different types of metadata unstructured mostly. And lastly what was interesting is we're starting to see a blurring of the lines between VOD and live. So it's no longer just you upload a video, it's also now I'm broadcasting a video. Or I'm broadcasting a video that the source is actually an on-demand video that I have on my computer and vice versa. So the consumers at the end of the day, the users who are using an application where video is important, they're starting to expect a lot of different capabilities, a lot of different experiences. And that complicates things on the back end. This is where CalterA is born. So back in 2008 it was mostly a website, right? YouTube meets Wikipedia. Let's do our collaborative video online. In 2009 we kind of sit back, look at how people use what we've built and go, there might be something bigger here. And really the next step forward was to build a platform completely open source that people can use to build their own experiences online. And together with all those assumptions and all those realizations that video is going to be massive at scale, it's going to be required in many different workflows and many different applications. People are going to need a platform to handle those workflows, to handle all those different solutions and put them together. And we put together this kind of manifest of what would the platform look like? What is important from things like horizontal scalability to the ability to handle things in different systems and being backward compatible and so on? 2010. Steve Jobs takes the floor and says, HTML5 is king, everybody should dump flash. Well who? That long live HTML5, now we have standards. Now people actually do care. Now the web is actually progressing faster and we're starting to see a lot more applications. Popcorn.js is born. We're starting to see a lot of interactivity on the web alongside video. It's not just video, it's video experiences. We're starting to see a lot of different embedded workflows with video and it's kind of the main asset. So in education for example, we're starting to see video being used for assessments. It's now actually counting against your grade as a student. During that year, those two years, we actually make two main launches, Kaltrav v2 and Kaltrav v3. It's now a massive platform, capable of handling a very large scale of video. The Internet Archive is actually starting to use Kaltrav v2 alongside Wikipedia. And adoption is growing very fast. 2012 till 2014, we're starting to see massive adoption all over. Many different use cases, many different environments from healthcare to education to surveillance even. And a lot of those years were around stabilizing the platform, creating all sorts of tools around it, working with the community to create all sorts of future proof workflows so that you'll be able to handle large scale video in the backend. That year is kind of ending with the release of a real-time package that we built based on Nginx. So everybody here knows Nginx. The beauty of it is being able to handle massive amounts of connections. And what's interesting, it's very useful for video. So we wrote a module that basically does real-time packaging of video formats, and we'll talk about this later in more detail, that's the year it was starting to release. So around 2015, we started launching standardized installation packages for Kaltrav, so launching of RPMs and devs, and then furthermore, we're starting to see significant growth. Those are actual installation numbers over the years, and that number is kind of hockey-sticking right now, and it's really exciting because it really shows that now, 10 years back when we looked at this world of online video, we kind of envisioned where video is going to be used massively across many different workflows, but we couldn't even begin to think about the size and the scale that this will take. So this was like a fast-food 10 years. Let's talk about a few notable FOSS projects that we have that you can check out. So first of all, obviously, the platform itself, you can install it from the platform packages, so the RPMs and devs, we also provide Chef and all sorts of other deployment capabilities. The source code for the backend, which is mostly PHP, is over here, and then there is the player, which we'll talk about in a bit as well. Really, what it's all about is handling everything in the world of video from a backend perspective. So handling multiple ingest points, handling video headscales in terms of your storage, in terms of your delivery, in terms of your processing, managing transcoding queues, all those kind of things. Then there's the Nginx Packager that we've talked about, so on-the-fly packaging for many different formats, including encryption and DRM, thumbnail capture in real-time, clipping in real-time, stitching in real-time, all sorts of things like subtitles and multi-audio tracks and so on. I really recommend checking this out. If you have a project that requires real-time packaging of video, it does both live and VOD today, and it can also handle the mix between them. So we talked before about lowering of the lines. Those type of technologies, this is what they do, right? They enable you to kind of treat video as a video, and it doesn't matter if it's live or if it's VOD. If it's VOD, it can become live. If it's live, it can become a VOD, and really the user shouldn't really care. MW Embed, a web-standard video player. The reason it's called, by the way, MW Embed, is because originally we wrote it for Wikipedia together with the MediaWiki Foundation. We actually had Calturians working in the Wikipedia offices embedding video into Wikipedia, and that's how MW Embed was born. Really the goal of it was to kind of create a unified standards player that doesn't really care what platform you're running on, what device you're running on. It will make sure the video works. Then it grew very, very significantly to hundreds of plugins and customizations, and there are all sorts of things for interactivity or for cue points and so on, all the way to custom looks and skins and so on. We also have what you see is what you get sort of player studio where you can go in and kind of change how things look like and add your own plugins and so on. We'll talk a bit about this in a bit. Lastly, there's a REST API client generator. We kind of took a different approach to REST API. There's a blog post here that I highly recommend writing. What's interesting about this, and I won't go over the list, but you can go on and check also the, there's a link to the GitHub repository inside the blog post, is really one of the massive challenges that we had to deal with massive scale. At the same time of having horizontal scalability, we had to deal with lots of limitations of enterprises, where people would want to deploy Caltura in an enterprise environment or in education, a university environment, where there's lots of firewall rules and complexities. How do you deal with all those things? We had to come up with a REST API that is simple just like REST, but at the same time circumvents a lot of the limitations that normal REST APIs have. In addition to that, there's also things that are specific to the world of online video, like handling large files, which most REST APIs don't deal with. Jam is another interesting application. It's actually a PHP module, so I highly recommend looking into if you guys have PHP applications. This basically runs on the PHP level and helps you do logging very easily and very quickly. So a little bit of information about what people do with Caltura today in over the years. I'm not sure why it hides a few lines here, but this is basically saying healthcare and this is saying government, I think. Really interestingly, so we saw thousands of deployments and so on literally all over the world. We're starting to see it used and embedded natively in many different applications, many different environments, but then this is really interesting. As time goes by, we actually started around here. So in the early days, the majority of deployments were distribution of video and things like advertising and so on. Nowadays, we're starting to see significant growth into areas like telemedicine or healthcare patient videos or all sorts of capabilities, workflows around places where video was not naturally used a lot and now it's actually almost transforming an entire industry. We're starting to see that a lot with compliance and safety in construction and so on. So really interesting. Culture is not just for video, but it really is. The focus was video and everything around it. So because video is so complex, generally it handles a lot of different assets around that video. So it's sort of an asset management system and people actually use it to manage all sorts of data type in various sizes and scales. All sorts of different companies all over the world from government to media companies and healthcare and so on. Really interesting slide. If you get to go on the presentation, you can actually click on them and see what they do. Where do people deploy? We're still seeing most of the people deploy on their own environments, but there's a significant growth into public clouds. So this is sort of the distribution that we're seeing today and it's also sort of an indication of who are the strong IS players, I guess. So what's next? I really have 50 seconds. So we've spent a lot of time building new developer tools. You can check them out at developer.catra.com. Lots of interactive recipe system. There's some innovation around how you guide people through REST APIs and so on. Native SDKs and new web standards players are actually... We just launched new native SDKs for the player. For native iOS and Android, we're actually starting to work on a new player. So kind of getting rid of a lot of the legacy of MediaWiki and the old days. Real-time, for two main reasons. One is because the lines are completely blurring again between live and VOD and real-time is definitely a big part of it. But also because there's a need to support large broadcaster and multi-participant experiences and we're actually seeing a lot of demand for that. Machine, so everything around sort of AI and the ability to kind of describe the video better and extract information from it. The future of YAR and all sorts of things around simplicity of deployment, simplicity of scaling and monitoring and so on. And definitely if you want to participate, if you want to share your thoughts, let us know. That's my last one. I'm sorry? Are the companies like Amazon or Google paying the software? No, so Caltra is completely open source. You need to repeat it with some... Oh yeah, sorry. Let me make sure that I understood your question. You asked if the large companies using Caltra are actually paying for Caltra. So Caltra is completely open source. For a GPL, you can just go ahead and install it yourself and all the logos that you've seen there are actually using it for free completely. But we also provide commercial services and we have a cloud of our own. So a lot of people don't want to install themselves. They just want to get a service and they simply buy from our hosted service. And so yes, there's a mix. Other questions? Yeah. What are you currently doing stitching directly to 160 video? So what do you plan to do and what can we do with the stitching and the 160 video? Yeah, it's a good question. So the question was basically around VR and what we plan to do around that, specifically around stitching. So today what we do in terms of stitching is basically the idea of clipping and stitching. So you get a long video and you can clip out of it or take multiple videos and then stitch them together. But that's not really stitching in the sense of VR where in VR you would get like a 360 video and flip it or trim it into like multiple quadrants or cubicles and so on. And then in terms of looking into the future, there are I guess two things. One is better support for 360 which we already released version one. Version two is going to clip a little bit more support. In the further future I guess in terms of VR specifically is also about interactivity, the ability to navigate within the video, the ability to communicate between the server so network optimizations and so on. So we'll see about that. This is kind of further along. Other questions? All right, thank you very much.