 Great, hi everyone. So I'll fill in my title, it's a bit long, so I run a number of services at AWS, I run Amazon Elastic Search Service, Amazon Elastic Cache which is Redis and Memcache as a service, and Amazon Neptune which is our new graph database. Today I'm going to talk about Open Source. Oops, as was mentioned before, I've been in Open Source for I think it's over 22 years, started to contribute in 1997. Main body of work was around PHP but also contributed to lots of frameworks, development tools, probably the contribution that I think is kind of the coolest that no one knows is I was actually one of the people who wrote the tab completion and the MySQL client. So if you ever used MySQL and used the command line client and had this really awesome command line tab completion that happened I think in 1998 and we were trying to test PHP with MySQL and it was really frustrating that we had to retype the select statements every single time and guess the table names and that was kind of a legacy that no one actually knows out. So today I'll talk a bit about Open Source at AWS and then I'll talk about how we think about it and how we think about keeping it open. So first of all, very important to state, we're a very, very customer obsessed company. When we innovate and when we do work, we don't look at our competition. We don't think of, you know, what's the coolest next technology and we definitely don't think of our short-term results. We have a saying at Amazon where really our goal is to build relationships with our customers that outlast any of us as individuals. So when we're making decisions, we're making decisions that really are thinking about the long-term and if you're not thinking about the long-term benefit for customers at Amazon, you basically don't get to make any decisions. And that's also how we think about Open Source and our customers really, really love Open Source. I mean, Open Source is what they run on AWS. And so a lot of the innovation that we're driving in Open Source comes directly from customer feedback. And sometimes we innovate by contributing to customer projects like Netflix's Spinnaker project. Sometimes we open source our own projects like Firecracker that we just open sourced. And in many cases, we go and contribute to open source that our customers use. We also use a lot of open source to build a platform. And, you know, what we see is we see a lot of our internal teams sometimes for just selfish reasons, not wanting to, you know, maintain technical debt. They want to make sure that all that code goes back upstream. And we actually have a very, very easy way for developers in our teams to contribute to Open Source. So that's sometimes a bit of a misnomer about Amazon, but it's actually quite easy to contribute to Open Source. We also care a lot about the communities because I think community is really what creates longevity for Open Source. Without community, Open Source really isn't existence. And then last but not least, you know, one of the things that I've seen in working in Open Source is co-quality typically goes up. You know, it's really fast to really easy to fix a customer problem in the cloud by using the control plane and doing something very fast, kind of a quick and dirty fix. But when you have to bring that code upstream, it gets reviewed by a lot more folks, security, right? There's a lot more scrutiny around the security. And so there's a lot of benefit, I think, in engaging in Open Source. So those are all things that are basically driving us towards doing more Open Source. We contribute to a lot of different projects. I basically, I had a tag cloud up here initially with kind of all the projects we contribute to, but it was a pretty ugly slide, so I just picked out a few projects. But you know, we're contributing to SED. We're contributing to Java, Nginx, Lucene, just a bunch of projects out there. And contributions from Amazon are definitely accelerating. We put a big focus also on keeping communities healthy. And actually, our Open Source team, a lot of them are in the crowd right now. That team is just growing every single year by quite a lot. And the core focus of that team is ready to make sure we make the investments also in keeping the communities open and keeping them, you know, keeping them energetic. So whether it's Linux Foundation, we just became a Platinum sponsor in the Patchy Software Foundation. We support the OSI. We sponsor a lot of conference. So OSCON, KubeCon. We're doing RedisCon next month. And then we also publish a lot in social media. We have our own Twitter handle in the Open Source group. We have about 20%. 20% of the content that we create is actually from guest posts. So it's from folks from outside of Amazon who are basically, you know, working with us to really make sure we get the Open Source message out. I was trying to, I was going to talk a bit about, you know, what are the different kinds of Open Source. And I felt that if I kind of really, well, you know, became very accurate, we would never get out of here. So this is a big oversimplification of Open Source. But I kind of simplified it into four buckets. You have individual authors, which was kind of me in 1997, where I thought it was really cool to kind of work in a language and, you know, worked with a partner of mine and started to, you know, work on that before the term Open Source was actually not coined at the time. So we didn't have a thing called Open Source. We didn't have a thing called GitHub. We just kind of started to work on this, you know, source code available stuff. Enterprise-driven Open Source. So things like, you know, Netflix projects. We just, you know, we just contributed Firecracker. And then, of course, Foundation-owned, you know, things like Kafka, Kubernetes. And then you also have the commercial OSS vendors, which Joseph talked about in the Valley creation there. And what's kind of interesting is that, you know, products actually move from one to another, right? Like Kafka started at LinkedIn, you know, was Open Source into the Apache Software Foundation. And then later on, also a commercial entity was built. Redis started from an individual, was then sponsored by VMware and Pivotal. So it kind of became an enterprise-driven project and then became a commercial-driven. So it actually gets very hard to categorize Open Source projects in any single bucket over time, because, you know, communities are vibrant and things change. But I think one thing that I'm here to talk about is there's been a lot of, you know, I'd say conversation from very few around how Open Source is having issues with kind of the new world and how that's kind of making it harder for Open Source to thrive. I think it's exactly the opposite. I mean, I can't remember a time where Open Source has been this vibrant. You know, when I started my company on Open Source back, I think it was in 2000, we went to enterprises and we couldn't even talk about our product value. 80% of the conversation was why Open Source was okay for them to use. Like, they were actually not willing to use Open Source. And then another 15% was why dynamic languages are okay, because dynamic languages were basically like these toy languages, you know, no strict typing, no enterprise application uses dynamic languages. And I kind of got the 5% opportunity to actually try and sell something. And, you know, now we're in an era where every enterprise is using Open Source and they're not just using, they're contributing Open Source, right? Service providers are contributing a lot. And also individuals, right? Individuals are basically building their resumes by going up to GitHub and starting to, you know, create Open Source software. Actually, most of that Open Source software we never hear of, because it's only a select few projects that are actually going mainstream. And even more than that, I don't think, as Joseph, you know, two speakers before me actually pointed out, we've never been in a time where there have been so much, you know, value creation around Open Source also for private companies, right? For 20 years, we're kind of seeking that business model around Open Source. And now where enterprises are adopting Open Source in a strategic way, it's actually becoming, you know, companies are really, really successful at monetizing that. And not monetizing that because they do, you know, make the license changes and so on. Actually, if you look at it, most of those companies actually got into their high growth trajectory in their hockey stick by having a very simple and sound Open Source business model. So how do we think about licensing and, you know, where do we feel at Amazon that, you know, Open Source kind of, you know, needs to stay open and what are some of the things that, you know, we think about? You know, first of all, you know, from our perspective is Open Source should be Open Source and proprietary software should be proprietary. And, you know, Open Source, you know, Open Source gets traction with the promise of community, with the promise of openness, with the promise for the end customer, right, that they have freedom to use that Open Source however they want, and they have the choice if they want to make it, right, to self-maintain in the long term. And so, those fundamental freedoms around Open Source I think are very important, you know, when you actually go out and build an Open Source project and get that broad traction with the project. So, you know, first thing is if you do Open Source, you do Open Source. If you do proprietary, you do proprietary. There's nothing wrong with a proprietary business model. You know, I have proprietary software. When I was building my company, we kept it very clearly separate. We had an Open Source project and we have proprietary. So, keeping that separate, in our opinion, is the right thing to do if you really want to foster a community and make sure that there is kind of a long, you know, you provide longevity to Open Source. The second thing is, you know, really making sure that when you have, you know, proprietary value add, which is a good, you know, is a good model right for many companies and it's absolutely something that, you know, has made companies successful, you know, just be very clear with your customers, you know, what they're paying for, what is Open Source, what they're paying for. So, there's great clarity for customers and what they're getting. The third thing is, you know, one of the core freedoms around Open Source is actually the ability, right, to view the code, to change the code, to distribute the code as long as it kind of adheres to whatever the license allows you to do, you know, making sure that innovation is enabled from our perspective is very, very important. And it's not just about the source code and the licensing, it's about how the project is managed, it's about, you know, a number of other things that basically matter there. And then I would say last but not least, when you put Open Source out there, right, you're making a certain promise, like we actually, you know, back in the day of PHP, we went out with a certain license and we wanted to add a more permissive license and we added that instead, we didn't take away. So, another thing I think is when you make a commitment to a community and to users, you know, make sure that, you know, if you want to, you know, modify your business model, do it in a way that doesn't change the rules on the Open Source in a retroactive manner. So, one I think the most important piece here is, you know, Open Source is a supply chain, right? Everyone is innovating on Open Source and when we think about individuals and organizations and startups that are trying to build awesome products really, really fast, right, they're basically building on Open Source most of the time. And what's cool about Open Source, and I think one of the things that I've always loved about Open Source is, you know, you build a product, right, you can get a bunch of pieces right from Open Source, right, and then really focus on your value add within that ecosystem, right? And, you know, taking an example, I'll talk about some specific pieces in the stack later on, right? Core supply chain pieces that we use today on a day-to-day basis are things like, you know, virtualization, Linux, of course, and Java, right? And I'll talk about elastic search a bit later on, but if you look at elastic search, right, what really made elastic search elastic search, it's Apache Lucene that came from the Apache Software Foundation, right, basically was created about 11 years before and is the key indexing under elastic search, also using NetE, Jackson, Lock4j. And so I think the thing to take away from here is, you know, Open Source feeds Open Source and I think a lot of the innovation we're seeing today, you know, is because there's this great virtuous cycle from being able to, you know, build on top of that. And then, you know, above that, you have ISVs, you have managed service provider, you have system integrators, you have a whole value chain, right, as was shown in one of the previous presentations of folks who can kind of get the benefit. And it's really not a zero-sum game because the reality is these market segments are super big and no one company can actually capture the value. Not Amazon, not anyone else, you know, it's hundreds of companies, even thousands that actually captured the value of these kind of communities. The problem happens when, you know, the supply chain breaks down and, you know, and that's a problem that, you know, when it happens retroactively, it's a real issue. So if you think about Java, right, Java came out in 1996, right, so folks had about 23 years of kind of having a pretty, you know, reasonable cadence around Java releases and so on. And then a few months ago, right, Oracle announced that they were going to change, you know, some of the Java licensing, the release cadence. Frankly, I don't even myself fully understand it because I think no one does. I think a lot of enterprises right now are trying to figure out what's going on. But we've basically gone into a six-month release cycle. So the security updates that most commercial applications get is about six months. And it's not quite clear, I think, to a lot of organizations right now on whether they have to pay or don't have to pay for a Java license and when they're going to have to pay. So frankly, I don't even know the exact details because I was reading, you know, both the Oracle content and other content that was out there. I think there's still a bit of a wait and see. But the big problem here is it creates a lot of uncertainty, right? We have 23 years of applications being built on Java. Uncertainty is never good in the supply chain. And that's something that, you know, we at Amazon kind of sort of was starting to impact our customers, right? Our customers were starting to get worried about the Java supply chain. And so one of the decisions that we made is when we see, you know, kind of the supply chain breakdowns, you know, for our customers, you know, we'll take a serious look. And if we think that there's certain things that we can help with to make sure that customers really have unfettered access right to what they were used to, then we're going to, you know, basically act and, you know, try and help out. And that's why we announced that Amazon Coretto a few months ago. Coretto is basically a distribution for Java that is free for everyone. You don't have to be an Amazon customer. And we are basically providing that long-term support for free. So we are going to be doing, you know, backporting security fixes, I think for Java 8 until June 2023, for Java 11 to, I think, August 2024. And so, you know, with Coretto, we've basically been helping out not only our customers, but I say to the overall community to make sure that we take that supply chain issue out of the system and we really help and make sure that enterprises can continue to bed on Java. Very important to note that a, you know, fundamental thing for us in creating Coretto is we didn't want to go and fork Java. You know, we think the OpenJDK project is, you know, great and fine. And so, you know, any backporting that we do, any work we do on Java, we actually do contribute back upstream to OpenJDK. So, you know, very, very important to note, you know, our goal here was not to bifurcate any community. We're contributing back to the main OpenJDK project. It's just making sure we create this distribution that customers can trust and they know exactly what they're getting. And they're getting the support that we think they should be getting over the long term. Another, you know, kind of breakdown that was in the news, I would say in the last 24 hours was Elasticsearch. And let me just say that I think, you know, Elasticsearch is awesome software and there's a lot of commercial success behind Elasticsearch and I think that's great. And as I said before, the market segment is super big. You know, it's in the middle of log analytics and machine generated data analytics. And there's just a huge ecosystem on top of this. And I think from our perspective, lots and lots of players in this ecosystem are going to do well. So our point of view here was really kind of a supply chain view. You know, we have a service called Elasticsearch. A lot of our customers also run Elasticsearch on EC2. They're running Elasticsearch elsewhere. And in June 2018, confusion started to enter the Elasticsearch world when a number of changes were basically made to how the code base was being managed. First of all, proprietary code started to be intermingled with open source code. Not only was it intermingled, it also started to become very difficult to actually understand what is an open source, what is an proprietary software, and that started to create a real issue. Code commits started to commingle open source and proprietary. So if you wanted to be an open source developer and not look at proprietary code, it actually became more and more difficult to do so. And you'd have to have some kind of cleanser on code commits that basically erases the proprietary code so you make sure you actually are only looking at open source software. So that started to become really challenging, you know, from our perspective and also from our customer's perspective. In addition, the key distribution point really moved to kind of the, you know, source code available proprietary version, which meant that most folks who came and downloaded it, and I think many of them probably thought they were getting open source, they were really getting proprietary software. And that's where the clarity comes in, right? I'm not saying proprietary software is bad. I'm just saying that it's a very, very important, right, to be extremely clear with your customers and with the community what's open source, what is proprietary. And then, you know, another piece of another challenge we had was innovation was actually shifting, you know, from open source to proprietary. We've got lots of customers using it. A lot of our customers love open source. They love the freedom to know that they can basically self manage over time if they need to, right? If they don't want to work with us directly, if they want to self manage on-premise, they want to have the assurance that they get the freedoms that they're used to from open source software. And so, you know, what we basically announced yesterday, similar to Coretto was basically a distribution for Elasticsearch. And that distribution has really a few key goals. Number one, just make it very easy for our customers who care. Not everyone cares, but the ones that care about open source to get a package that is guaranteed to be Apache 2 licensed and is completely open source, and they can do whatever they want with it. And that includes running it on competitive clouds, running it on premises, running it on their laptops, right? This is true open source from our perspective. It is not Amazon open source, it's community open source. Second, as we saw innovation kind of slowed down on the open source side, we felt it was really important to start also innovating on open source. And so we've added capabilities there, such as security, SQL, alerting, and, you know, our goal is to continue to innovate in open source and give that code, you know, back to the community and make sure that whoever is interested in this kind of distribution can get that kind of unfettered license distribution that gives them long-term control of their environment. And then last but not least, the goal here is to, you know, build a community. We've had customers like Netflix and Expedia kind of, you know, join us when we announce this. But we've seen quite a lot of interest, you know, just in the past 24 hours of, you know, other customers and partners that have potential interest in this. And so our goal is to truly build a meritocracy here about a community driven project where it's not only Amazon, but it's really others who are contributing to this project. And we're making it very easy. We've even eliminated the requirement for contributor license agreement for a lot of the smaller patches. It's basically a click-through approval that they're contributing under Apache 2. And, you know, we're going to continue to invest in this and make it available. And then some of those features are going to find their way back to our managed service. But, of course, you know, anyone is welcome to take this distribution and innovate on top of it however they see fit in any environment that they want to run it in. So, you know, we're open for contributions. By the way, we are partnering a lot right now, both in the CNCF, you know, around a lot of different communities on contributing with partners and even competitors around open source. And the same is true for here. You know, we're welcoming anyone to come and join us. You know, if you're interested in an open distro for elastic search, you know, talk to us. And, you know, our goal is just to make this a good, credible project for the community. Thank you.