 All right, so I guess this is the annual tradition. I come up and get everyone up to speed on the latest of where the project's at and kind of pick at a few of the features that the project's been working on. So let's dive into it. All right, so first I just wanted to say that this is the fourth year of MesosCon. Which is really impressive. Who was at the 2014 Chicago MesosCon? Anybody? A couple people. All right, who's been to all of them? Two, three, Josh. All of them. That's amazing. And it's been really fun to see to see the MesosCons go from Chicago to Seattle to Denver and now this year in Los Angeles. It's great to keep the tradition going. Again, this couldn't be possible without our sponsors, all of them. You're what helped make MesosCon happen. MesosCon is a community event and without the sponsors we wouldn't be able to do what we do. So a big thank you to you guys. We've got one more MesosCon coming up. It's coming up in Prague. It's our MesosCon Europe. It's October 25th through the 27th. There's a great collection of presentations there. We even had some presentations that we really wanted to accept here in North America that we ended up moving over to the Prague event just because we had so many great selections. So please go check it out. It's the last MesosCon of the year. We had one earlier this year in Asia. It was our second MesosCon Asia and it was really, really impressive. We went from having really two kind of user talks at our very first MesosCon Asia to over 14. So we had 14 different end users in Asia, specifically in China. They were using Mesos in really serious capacity and they all gave presentations about what they were doing. It was really, really impressive and humbling to be there and hear from super large organizations like China Unicom and China Mobile talking about how they're actually taking advantage of this technology internally. It was really, really fun. Okay. A few additions to this year we decided to mix up MesosCon a little bit. So we have a few new additions. One is we decided to introduce tracks. We recognized that a lot of organizations, sorry, a lot of people that were coming, they were coming for specific interests. And as we were sort of surveying what was happening in the Mesos ecosystem, we thought it would be much better to really split the conference up to tracks. So there's a couple of tracks that are on two days. That's the SMAC GPUs and analytics track. You can hear from lots of people that are actually in the field doing SMAC with Mesos, using GPUs, running lots of analytics. There's two whole days of that. So tracks on both days. There's an ops track. We're recovering everything from storage, networking, monitoring, security. Again, there's lots of talks around that. So we split that into two days. There's a DevOps track, which is a bit more focused on microservices, APIs, continuous integration, continuous delivery. And that we're just going to be doing today on Thursday. And then tomorrow there's some tracks that are just tomorrow. There's a Mesos frameworks track. So anybody who's using frameworks, hacking on frameworks, want to build their own frameworks. There's a bunch of presentations around that. And then there's a Mesos internals track, which is really some of the nuts and bolts of some of the new features that are being built in Mesos. Some of the details about how those are being built and how they can be consumed. We also decided to add some longer sessions. We've gotten some feedback that in some presentations, people wanted to get even more hands on. And they wanted to sometimes even be able to use their laptops and play around with things directly in the sessions. So we introduced something new this year that we're really excited about. We called it MesosCon University. They're 90-minute tutorials. They're interspersed without the presentations today. So there's three of them. There's my application is ready for production now what. I think you're going to go through everything from using health tracks to doing some scaling to doing monitoring. There's a bunch of different things you're going to do there. At 2 p.m., there's one all around security, boot scrapping secure Mesos clusters. And then at 4 p.m., there's one around building a stateful DCS service using the SDK from Mesosphere. So again, they're about 90-minute sessions. They might go a little bit over. You should bring your laptops. They'll be fun. And it's going to be a presentation, but there's also going to be a time for you to actually get involved and write some code and get some stuff actually running. And then the last addition was we wanted to put together these town halls. And the basic idea of the town halls was just a place. We noticed as big community gatherings at these Mesos cons, we typically didn't have a great place for people to just get together and talk about the projects always. They were going to presentations and hearing about what was happening, but there wasn't as much community discussion as we really wanted. So we decided this year to introduce town halls. They're 7 p.m. tonight after the booth crawl. There'll be beer. There'll be some snacks. There's going to be three of them, one on Apache Mesos itself, which is also covering some networking and storage, one around DCOS and run around Marathon and Chronos. And so please come to those later and hang out with peers and talk about the technologies and discuss anything and everything. So I wanted to give a big thanks to the speakers. We've got a ton of different speakers this year. We just highlighted some of the new speakers. We have some brand new speakers that have never actually spoken to MesosCon before, and I thought it'd be fun to highlight them. There's lots of people that continue to get involved in the ecosystem, that can continue to get involved in the product, a bunch of new speakers, and I think that's really great. If I missed anybody, I apologize. We tried to make sure we were just capturing the new ones, but really exciting to see that. All right, now for the last couple of minutes, I'll just give a quick overview of the state of the project, where we're at, and then we'll kick off with the rest of MesosCon. We've done a ton of releases. There's been 12 releases since last year's MesosCon. Some of you might remember that last year, the big announcement was our 1.0 release. We did that, and since then, we've done a 1.1, a 1.2, and a 1.3, so three minor releases since that, and nine patch releases total. So the community's been really busy working on those patch releases and getting those shipped. We've added 13 Committer and PMC members since last year at MesosCon. It's a really, really huge number across tons of organizations, from IBM to Apple. There's been a lot of really great contributions, and we've been excited to add those members to the group. When I was asking the community, when I was asking some folks for other things that they're proud of seeing, this is one of the things they mentioned. They're excited that we continue to have a growth in our active members on Slack, people that are joining in the conversation and the discussion, both from a development perspective, as well as just a user and operations perspective. So it's great to see that growing as well. So I asked a bunch of folks in the community. I said, all right, every MesosCon, I come up here and I kind of talk about some of the new and noteworthy stuff, and what should I talk about? And I got this giant list, and this isn't even all the things that I actually heard. Of course, I don't have time to go through all these things, but it's been really impressive to see the community really doing a ton of work in the last year, and I thought I'd just highlight a couple of them that I think are especially interesting and I wanted to share with everyone this morning. The first couple of ones are all around containerization. So containerization in Mesos has been something that really we started a long, long time ago. When the project first started in 2009, we actually used Solaris Zones and Projects to do our containerization, and we pretty quickly used LXC, which transitioned from there to using C groups and namespaces. We eventually added support for Docker, and in the last couple of years, what we've done is we've taken some of that original C groups and namespaces work that we did and packaged it up into what we call the UCR, the Universal Container Runtime. We had this predicament with the project, which is we were never going to be able to really get rid of all the containerization work that we'd done in the past, because we had really large-scale users who wanted to continue to use that containerization tech that we built, and so we tried to have to figure out what's the best thing for them when they wanted to do things like, say, run Docker images. So Universal Container Runtime, it's pretty cool. It's the original containerization code that was in Mesos, but now you can run containers with a Docker image, with an App-C image, there's code out there for an OCI image, or you can run containers as was before the day as a Docker, without an image at all whatsoever. So this is pretty fun. For organizations that aren't necessarily moving to something like Docker, but still want to get the benefits of containers and container orchestration, they can directly leverage the UCR to do that, and there's a lot of organizations that are doing that. They're just taking their zip files, or they're taking their tar balls, or taking other things, and they're launching their containers. Okay, one of the really cool things that we were able to do because we have this containerization stuff built in directly to Mesos is stuff around nested containerization. So we actually built arbitrary nested containerization. What's that let you do? It lets you create arbitrary nested containers, an outer container, an inner container, an inner container. You can launch a container that starts from a zip file. It can actually then launch a container which comes from a Docker image. That might potentially launch a container which comes from an OCI image, and there's lots of really, really great applications to this. The two big ones that we're excited about are things like Jenkins. When you want to run Jenkins, Jenkins itself might be launching containers. You want those containers to be contained. If Jenkins happens to die, you don't want the orphan containers to be running around on your machine. And the other one which we're excited about, which you'll hear a little bit more about later, is for running Kubernetes on top of DCS and Mesos, where again we can do that in a way where our top level containers don't actually need to be privileged. They can just use capabilities. So it's been really cool to see this tech evolve in Mesos. As part of that, we added support for container attach and container exec. There's lots of interesting stuff there, but this is really then driving a bunch of work that we're excited about, which is standalone containers, which is just going to be the ability to run containers not through the typical Mesos APIs, but just any standalone containers, especially demons, which has been a huge request from the community for a long time is to better support demons. So lots of great work in containerization that's really building up to a bunch of these features. The other area that I wanted to touch on briefly related to containerization is networking and storage. This is one of the areas where I'm really, really proud of the Mesos community because they've really taken a leadership effort along with a bunch of organizations around this. The two big focus areas for us have been around CNI, the container networking interface, and CSI, the container storage interface. And I just wanted to give a shout out to GU and JDF for doing a great job really helping to drive forward the container storage interface. This is super tough work in the industry to actually make changes like this across vendors and users, and it's been really, really impressive and fun to have the project be writing in the midst of it all. Okay, one of the things that I'm especially excited about when it comes to the storage work that we're doing is we're finally building out support for what we've called local and external resource providers. These are third party resource providers that can expose new kinds of resources, can deal with the allocation of those resources, the quota of those resources, so forth and so on. We've talked about this for a long time in the project and it's becoming a reality specifically around storage. So what does this really mean for you guys? It means better integration with third party storage. It means better integration with third party networking and you'll start to see that pretty early on as the work around CSIs is really enabling LVM local storage as well as network attached storage in a much cleaner way than we've been able to do it in the past. Okay, the last one I really wanted to highlight because I think it's really great to see is all the work we've been doing around multi-tenancy. I looked last year at the slides that I presented and that's when we first talked about quota. So we just introduced quota last year and this year we've improved that to not only have quota but have hierarchical quota, hierarchical fair share as well as this concept we call hierarchical reservations or really reservation refinement. So this is really, really powerful stuff. Most organizations are hierarchical in nature and by adding this we've really enabled organizations to delegate these things, to have isolated visibility into these things. It's all authorized and soon frameworks like Marathon and I hope others will actually be able to take advantage of these primitives themselves. All right, so I kind of wanted to end one last slide. I hadn't really done this in the past but I thought it'd be fun to just make some predictions about perhaps what we'd see in 2018. These are a little bit random but perhaps they're also my pet peeves. I'd love to see this stuff done. One is we're really excited to see more work around preemption. We've introduced the concept of revocable resources in the past into MESOS but we're not using it as pervasively as we really wanted to. So I'm excited to see in 2018 I think you'll see revocable resources being sort of front and center feature in MESOS. I think we'll probably see the end to ZooKeeper which I think, yeah, okay, there we go, yeah. There's been a bunch of work around either replacing it with something else and I think actually in the future what we'll see is just nothing. You'll just be able to launch MESOS without any of those third party components and it's just gonna be one less component you have to manage. And I think the last one is there's been a lot of talk from a bunch of people about extending the agents, the launching and containerization to not just work with images but also work with virtual machines. So there's a bunch of organizations that are looking to consolidate some of their virtual machine infrastructure with their containerization infrastructure and I think this is something that you'll actually see. So that's my real quick kickoff, really excited about as Peter said a bunch of amazing presentations that we have lined up for the next two days, a bunch of really, really great contents. We're really excited to have you all here. Have a great MESOSCon and hope to see you all in Prague.