 Right, so this morning we're gonna change things a little bit in terms of our speaker lineup. I don't know how many of you get to attend conferences other than LCA, but something I kind of enjoy and something I wanted to throw into the program was a mix of speakers one morning, keeping it short, sharp, interesting, hopefully relevant. And I'm really pleased to look at some of the people from the industry that we've got here this week. And so I tapped a few of them on the shoulder and said, look, can we do this? Can we make this happen? And I'm pleased that they said yes. So, we have Cooper Lees from Facebook, Carol Smith from Google, and Mark McLaughlin from Red Hat. So, if we're ready, I'd like to introduce Cooper Lees from Facebook. Good morning, everybody. How's the heads? Yeah, don't worry, I feel the same. But I've got experience with hangovers. So, we should be fine. Not with microphones. So, how you doing? Good, good. Yeah, we should be good. All right. Alrighty. So, today, why we get this happening, I'll start talking, because I know we're limited on time. I'm going to talk to you about Facebook and our open source initiative at Facebook. Just a quick show of hands. Who here knows that Facebook open source is a lot of code? That's pretty cool. That's a lot. I'd say it's about 60 to 70%. And who here has actually got some Facebook quality, bug-free code running in their production environment? Who's running ButterFS? Who's using Mercurial? There's plenty of projects. So, we, and one of the cool things working at Facebook are encouraged to open source things. We've tried to improve when we go back to some of the early projects we threw over the fence. We literally did that. And we've now tried to focus on doing the full ecosystem of an open source project. So, let's get numbers, because nerds love numbers. We have hundreds of GitHub repos, and GitHub is our primary source of where we throw our open source code to. We have thousands of contributors from all across the world. Weird speaking, Australians, such as myself, all the way up through North America, everywhere. And we get a lot more collaboration and involvement across all our repos these days. We love graphs, so let's look back at 2012, where we first sort of started open sourcing many of our projects. And we've sort of grown all the way up to today. You'll see that big dip at the top. That's us throwing out some old projects that we no longer maintain or look after. And we think it's a disservice to leave that code up there if we're not going to work on it. You see the same kind of trends with our commits, naturally. The number of total commits in all our repos. And the commits per month. I always thought this would be a weird statistic, but you can see the trend up. Where you see the big spikes is where we fork projects. An example would be MySQL into WebScale SQL. Would be one that would cause a big growth there and things like that. One of my favorite stats looking over the numbers with James Pierce, who's our open source sort of liaison, is that our pull requests have grown. So I thank you guys, whoever submitted a pull request to anywhere at Facebook, to any Facebook code base, thank you. And this is one of the main reasons why we do it. Yes, we get to share our cool code that we do write, but we also get value from you guys sending it back. So please keep it up. We want to be good citizens. We know that we were not perfect when we first open sourced our code, and I'm sure you've probably cursed our name before. But we're now trying to do a better job. We've ever made an organization or a collaboration called todo, todo.org, where it's with other leading industry companies to try and work together and do it better for all you. So let's talk about some of our top projects. Is anyone running any of these? So React is a JavaScript UI framework that we use extensively throughout all of Facebook in dub, dub, dub and other things. HHVM is one of our most popular open-source projects. It is the PHP virtual machine runtime that has made PHP quicker for not only us, but also recently Wikipedia, which I believe was two to three X times quicker running on HHVM as opposed to mod PHP with Apache. The async display kit is, as it says, an asynchronous display kit for your iOS apps. It used extensively throughout all our iOS apps. And one of my favourites is Presto. And Presto is an SQL-like interface that sits on top of large data stores so that you can query it and pull the data out that you want. So you can put that on top of RocksDB, for example. We also have a culture now of upstreaming things. I, myself, have upstreamed some code. Don't yell at me if you get some bugs. But some examples there are Chahire, which is a Go BitTorrent tracker that we use extensively within Facebook. LibTorrent, I've put some fixes up, predominantly for IPv6, now that we have v6-only clusters. We have many core Utils developers. We have two of the leading ButterFS developers. We do a lot on grep. There's like a speed increase of some significant number that we recently released into grep. I'm sure you've all used grep. And then Mercurial. We're really hammering out and scaling Mercurial for our needs and your law benefit from us making Mercurial faster and easier to use. So code.facebook.com is our hub, our blog, and talks about all the things that we're open sourcing and what it does, and also links you to all the code repositories. And github.com.com.facebook will get you straight to all our code. So let's go a bit more in-depth to the area of Facebook iWork. I work in the NetEngine department. We have a large network. I was larger. Google was large. And we have recently started dabbing and running our own Linux-based top of rack switches. So I'm going to talk a bit more about that. OCP is the Open Compute Project. Who here has heard of the Open Compute Project? Excellent. That's a big show of hands again. The Open Compute Project is where we're open sourcing our complete data center ecosystem so that you guys can all benefit. And I was actually talking to someone last night at the dinner who's using our rack design, our service design, and possibly the switches as we move forward. Fboss, as we're wearing the shirt, and if you have any questions about Fboss, the rest of the team is wearing the same shirt today. Is our open source switched software? It's not open source yet, but it will be. We're looking to do that. We want to make it bug-free first. So why are we doing this? What do we want from our open source network devices? So think about when you deal with your network vendor today. It's pain-free. There's never a bug. It works so well. Said no one ever. So we want to try and break that apart and disaggregate the software from the hardware and make the switches more open and friendly and more automatable. We want programmability, we want openness, and we want to treat our switches like we do our servers. We have a vastly larger cluster of servers throughout the Facebook infrastructure than switches and they're a lot more automated. So we want to do that to our switches. So how are we going to do that? This is the first Facebook-built data center in Primeville, Oregon. Right now, it'd be snowing there and a lot colder than that photo. But this is Primeville, Oregon. And when we designed all this, we actually did a lot of work first and designed all this and had all this design that we benefit from, but we don't make money from data center design. We don't make money from the rack design. We don't make money from our standard server design. So this server that I'm showing here is our standard web server. When you get a www.facebook.com or open your mobile app, these are the things you're hitting. There's many of them. These are our racks. So this is the sort of open compute rack that you'll see and the PDUs to the side. And all these designs with the complete data center infrastructure are open source for you to use. And I was so excited last night to hear that a data center being built here in New Zealand is using these designs. The open compute project. It's been really successful, just like even hearing as further strengthened that. Why did we do open compute? The big thing is saved dollars. We saved a lot. We have more control over all our infrastructure now. Yes, we have to have and control more of the pipeline. We have complete control about changing designs, working for us and scaling. And then to get the benefits from other people using it. So we sort of group open compute into multiple different facets. We have sort of the normal service. So I showed you our web server. We like to refer to them as webbies. And then we have like all the different parts. So we've open sourced our database servers. We've open sourced our storage servers where we store all your photos, cats, pictures, selfies, all the important things of life. And the last piece that where we're really adding in is going to be the networking part. And I'm really excited to be working on that and working on putting in our new switches. As I said a bit earlier, we want to just aggregate and pretty much fully open our network and automate it. I don't want to configure a switch ever again if I can avoid it. Somehow I find that hard, but at least we can automate it. So where are we at with the networking project in open compute? We are here today. We have open source switch, which I'll introduce in a bit. It's called wedge. And then we have operating systems from different vendors already out there for people to use today. So what is an open compute top of a rack switch? At its simplest form, this is it. We have compute, memory, RAM, storage, running an operating system. And then we talk over hardware to the hardware switching AC, which is just merchant silicon, which all the big vendors use as well. So what components have we had to work on to get the switch up and working? Well, first of all, we need an installer, naturally. And this is just an x86 server. So it's the same pixie boot install method that you'll use on all your Linux machines today. So you can use the same tool that you use to install OSs. You get an operating system. So our switch is running CentOS, funnily enough, the Facebook version of CentOS. And then on top of that, we have an API to talk down to the AC and program it. So to put the switch port into its VLAN, get some routable interfaces, talk and get them usable and forwarding frames and packets. And then on top of that, you need the routing control, the brain to actually work out the topology, talk to other network devices and make sure it's configuring the ports correct. So let's treat our switches as servers. Our switches actually run Chef as well and have cookbooks that configure all the NTP and do all that. And that is awesome that I don't have to worry about that anymore. So let's go to the hardware, wedge. Has anyone sort of seen wedge or using the open compute switches? I found one guy last night. Anyone else using white box switches? It is very new. So wedge is 16 ports of 40 gigi. What we're also doing to put it into existing racks with 10 gig next is using breakout DAC cables directly attached copper so that we can have four 10 gigs per 40 gig and that's how we can fill the whole rack up. It's the Facebook blue, matches this shirt and that's very important with nice blue LEDs. And we've actually wrote some sweet code to do fancy things with the LEDs. That's the most important part. I remember we were sitting in the lab that we had a temporary lab that we had near where we sit at Facebook one night and we're all sitting there watching the lights and one of the HR people that sits nearby walks past and goes, what are you doing? We're all sitting there staring at the LEDs and he goes, is that why we pay you all the money for them? We're like, yeah, this is the coolest part. You can't beat blue LEDs. So that's wedge. It's pretty, it works. I've seen it work. It's folding packets in our data center today. So what's in it? Basically you'll see here, as I said, it's just our OCP server plugs into the board and basically programs the merchant silicon. In this case, it's Broadcom and all the switch ports and the LEDs. Epos. We need to control a demon or a routing demon to both program the switch ports and do the routing. So today it's a C++11 demon that sits there, starts up like any other normal Linux demon, runs and will talk via the API to the switches. We'd like to disaggregate the protocols out of the controlling demon and talk thrift. We're using the thrift RPC API or calls to talk and communicate. We're trying to get away with the traditional CLIs and just have a config file and automation change config as you need. We're looking forward to open sourcing it and it is a priority to get done. So if we can help you route and switch in the future. So let's treat our switches as servers and move forward to open sourcing more of our stuff. That's sort of what we're working on at Facebook in the open source realm and thanks. Right, we're gonna do any questions if we've got time at the end or you can catch up with all the participants in this afterwards or in the corridor. Okay, so we just wanna keep things running. So now I've got Carol Smith from the Google Summer of Code program and the open source program at Google. Yeah, that was very nice. My name is Carol Smith. I work in the open source programs office at Google and I'm here today to talk to you about the program that I manage Google Summer of Code. For those of you who aren't familiar, Google Summer of Code is not advancing my slides. Hold on just a second. Let's try just representing how does that sound? There we go. Google Summer of Code is an online international program that is designed to encourage university student participation in open source software development and it started with the idea that university students should get experience that is related to the things that they're studying in school outside of university and it would help them in their careers after they left school. So the program provides a framework for students to work with an open source organization of their choosing and pays them a stipend in exchange for that development so that they can focus on the work that they're doing for the organization. Students get exposure to real world software scenarios like etiquette on a mailing list, working with developers across time zones and using version control. And this isn't just an experience that's useful for students who are majoring in computer science. These are skills that are applicable to all sorts of university students who are majoring in all sorts of different things. They're just useful life skills. So the students also get references in contacts within the open source organizations that they work with. These contacts often last years, if not lifetimes, people make friendships and they also get references that they can use while they try to network in their field and also after they leave university and are looking for their career. Also many of the students who participate in Google Summer of Code are new to open source and so this also helps bring in new developers and new lifeblood into the communities that they work with so that our open source organizations are continuing to have thriving, vibrant communities long after Google Summer of Code. And Google Summer of Code gets more open source code created for everyone. All of the successful student projects at the end of the year are posted to our program website and are available for the public to use and access however they so choose within the licensing requirements of the code. So the general framework is that early in the year and this year it's going to be in February, which is coming up soon so pay attention if you think you might be interested in this. Early in the year mentoring organizations apply to Google, submit their applications to participate and then Google reviews those applications and chooses the organizations that they would like to participate in the program. That's gonna be in March of this year. And then students submit their project proposals to the organizations that they would like to participate with. And they can submit up to five proposals to potentially five organizations or one organization. It's up to their choosing how they distribute it. So they have a wide variety of organizations to choose from and can submit a bunch of different kinds of project proposals, some of which might be more applicable to the organization that they're applying for depending on what projects they're interested in. And then the organizations will choose, review the applications and choose which students they'd like to participate. They also pair them with a mentor that they would like to help them through the term of the program so that the students always have someone that they can reach out to and ask questions of and someone who can really help them through the whole process. We announce which students are accepted in April. And I just want to take a second because this is the 11th year of the program. We celebrated our 10th anniversary last year and we've learned some important things about what makes Google Summer of Code so successful and what has made the program so long-lived. And one of those things is this student-mental relationship that I was talking about. It's an incredibly integral part of why the program succeeds every year. Having someone who you can ask questions of and who can really help you through the open-source community that you're working with is really an important part of the whole process and of the program's success. So once the students are announced and accepted, the students work on a project for about a four-month time period and they execute two milestones that they laid out in that original project proposal for their organizations. And the students and the mentors go through two evaluations, a midterm and a final. And this isn't just the mentors evaluating the students on their project progress. It's also the students evaluating the mentors and their organizations on how well they're doing as a community and mentoring that student. And we take a lot of that feedback into consideration in the following years of the program to make sure that the organizations that are mentoring the students are doing a good job and are continuing to foster a really open community. At the end of the term, in addition to passing a final evaluation, the students also submit their entire project to our program website. And again, that becomes available for anyone who's interested in participating to use. So as I said, 2014 was the 10th anniversary of the program and thank you. And thank you for the applause because my next slide is going to be that we really want to say a really profound thank you. There's probably some people sitting in this room who've participated in Google Summer of Code over the last 10 years as students or mentors or organization administrators and we literally could not have done it without you. So thank you, give yourselves a round of applause. You guys are what has made this program as successful as it has been. So we made some changes to the program in 2014 in celebration of the 10th anniversary and those are going to be staying in place for 2015. It's a program and I want to talk about a few of those. One thing is that we raised the student stipend from 5,000 to 5,500 US dollars for a student who successfully completes the whole program and so that's gonna be staying in place for 2015 and hopefully it keeps the program competitive with other internship and other opportunities that the students might have when they're looking around at university. Another is that we accepted more Google Summer of Code students than we ever have before, 1,307 in all and we're hoping to accept a similar number in 2015. And again, this is an international program so these students are literally from all over the world. Over the past 10 years, we have had 8,616 students participate in Summer of Code. Please give them all a round of applause. And we also accepted more mentoring organizations in 2014 than we ever have before, 190 in all and again, we hope to accept a similar number this year as well. In the last 10 years, we've seen the most students from the United States, India and Germany but we are seeing a trend where we're having more and more students from underrepresented and developing nations participate and we're really proud of this and we hope that this trend continues. For the last three years in a row, we've had more students from India than we have from the United States and I hope that it continues to grow that we see more students from new countries where we've never had participants before and hopefully you guys can help us get the word out to those students in those developing nations, those underrepresented nations so that we can continue to see a really diverse community participating in Summer of Code every year. And lest you think that Summer of Code is not represented in the Southern Hemisphere, we have had 28 students participate from New Zealand and 83 students participate from Australia over the last 10 years. Yay! We estimate that over the last 10 years there have been 50 million lines of code produced by Google Summer of Code students. And also Google Summer of Code, as I said, isn't just for computer science students. We've had students who have been majoring in cartography, linguistics, music, film, business. There's all kinds of majors represented in this program and these are skills that help students no matter what they're majoring in. It's not just specific to computer science. It's also not just an undergraduate program. We have students who are graduates and PhDs who participate every year as well. If you think that you might be interested in the program, I would encourage you to participate and we'd love to have you. And our applications for mentoring organizations are gonna be coming up on the 9th of February and they run until the 20th of February. If you represent a mentoring organization that you think might be interested in applying, you should speak to your community and start getting together an application. We also post all of the questions that we ask on our application for mentoring organizations on our website so that you can just start preparing your application now with your fellow folks in your community so that you are all prepared for February 9th. And we're gonna be announcing which organizations are participating on March 2nd. And then if you are a student who might be interested in participating this year, we are gonna start accepting student applications on March 16th and those run until the 27th. If you are a student who is interested in participating, I would encourage you not to wait until the student applications open. I would encourage you to get involved with an open source organization now and start volunteering and contributing. It will only make your applications stronger. We're gonna announce which students are participating in the program on April 27th. And then we have a community bonding period that runs for a month from April 27th to May 25th. And this is another one of the key aspects of why we feel Google Summer of Code has been so successful is because it provides a month for students to get familiar with the code base they're gonna be working on and also to learn any skills that they need to know for their project. And it also lets them get to know their mentor and their community and makes them feel more welcome, more at home and more comfortable with the work that they're going to be doing over Google Summer of Code. They're not actually expected to start coding until May 25th and so again, they have some time to get adjusted and get familiar with what they're gonna be working on. We have, as I said, mentor evaluations that start on June 26th and run until July 3rd. This is actually a change from last year's program. We're gonna be running the evaluations for a full week as opposed to five days to give people an extra few days to submit their evaluations. And like I said, this isn't just the mentors evaluating the students progress, it's also the students evaluating the mentors. And we do pay close attention to what the students say about their organizations and we take it into consideration. Final evaluations are the 21st of August to the 28th of August and all the students who pass their final evaluations are also expected to submit their code by September 25th. And I just wanna take a moment to say that if you're thinking about contributing, GSOC is a really rewarding experience and we've heard from mentors over and over again that this is a really wonderful program and if you're on the fence, I would encourage you to participate. Students as well have formed lasting relationships and bonds that last year's, if not lifetimes and it's a wonderful program that lots of people have benefited from all over the world. Just some links for you guys if you're interested, google-malange.com is where we administer the whole program from and I have a short link to our discussion list if you'd like to just see what kinds of questions are being asked or contribute to the discussion. That's a short link for our mailing list you can subscribe to and then we have two manuals that I would encourage you to read if you're a potential student or a potential mentor or a potential org admin. These are written by students for students and by mentors for mentors and they have really hands-on advice and interesting knowledge for you guys to make use of. Also, I just wanted to say one note about the conference. A lot of the AV volunteers who are here were funded by GSOC, which is awesome as well, so. Thank you guys. Thank you Carol. And as Carol mentioned, yeah, we've got a team, several of whom have come in from India who are working on the AV software for this and future conferences and they're working with Tim and the team around developing the HD platform. Hopefully we'll be using a future LCA, so I mean it's great to see programs like this benefiting conferences like ours and moving things into the future. So coming up next, we've got Mark McLaughlin from Red Hat. So we've heard it from Facebook about some of the cool things they're doing in the data center, particularly around hardware in the data center. Now let's have a look at how software is impacting the future of the data center. Hey, morning everyone. I'm not here to talk about GNOME, that's just my desktop background, although it did work on GNOME for quite a while and still a big fan of GNOME. Here I am. So my name is Mark and I work at Red Hat on OpenStack and I'm here to talk to you today, I guess from Red Hat's perspective, how we see open source really changing what's going on in the data center these days. So times are changing, right? Hopefully we're all seeing it, we're all aware of it. We're aware of things like infrastructure as a service, platform as a service, new application architectures like microservices, how people are aiming for scale out architectures and their applications. We're aware of things like continuous deployment, agile and DevOps. These are all part of the same team, the same trend, the same kind of desire on the part of people who want to have ideas and want to bring them to the market, bring them to users as fast as possible, get feedback and really iterate and really learn from their users. So this is our new reality, right? This is the world we live in. You might notice one morning your phone, your Google Hangout app, for example, has completely changed one morning. This is the world we live in, but it's evolving like crazy. And so like often in this space, the behemoths, right? The Amazons, the Googles, the Facebooks and the Microsofts, they're leading the way, they're showing a lot of the trends here, there are a lot of the innovations coming from there and they're building it on open source. And I think that's really positive and there's a lot of opportunities there. And we're seeing, we heard from like Cooper and Carol about a lot of the work they're doing in the open source world and the benefits we're getting from the work they're doing there. Because we don't want it to be like this, right? This stuff, this new way of delivering applications, it's a competitive advantage, right? It's, if you can master this, if you can master these ways of architecting your applications, if you can master these ways of operating and delivering your applications, it's a real competitive advantage and everybody wants a piece of this, right? Even telcos. So this has really surprised me lately and I'm just learning a lot about it myself. For me, I guess I've worked in the open source world for 10, 15 years now and my really only understanding of the telco space up to now has been we're special, right? Telcos, any conversation, it's about how special the telcos are, how different the use cases are there, how when you pick up the phone, the dial tone has to work, lives depend on it, demands for high availability that no other sector has, demands for deterministic performance that no other sector has. It can be quite an off-putting conversation sometimes but telcos these days are really squeezed and really scrambling, right? If you think about how much you used SMS five years ago versus now you're using Hangouts or WhatsApp or Snapchat or whatever it is, telcos really are, the world is changing very quickly for them and they're learning that they have to adapt quickly. So telcos need this agility, they need this responsiveness just as much as anyone else does. But if you look at the current telco data center, the reality of telcos' lives right now, these data centers are filled with big expensive proprietary boxes and it's not just networking forward in gear, it's not just the mobile packet core, it's not just all of that kind of stuff you think about. It's basically all of telcos applications, all of their ways of delivering value to others users, they're all locked in these big expensive boxes. And these applications, they take years to build, they take years to phase out and they're really difficult to scale and this is the complete opposite of the agility and responsiveness that telcos need in this world. And so to me, that's how I imagine telcos these days, just pure frustration, they can't respond fast enough and they need a new way of doing things. So you might have heard of this new buzzword acronym that's going around at the moment called Network Functions Virtualization. It took me a little while to understand what this was all about and how it fits in the broader scheme of things. But it's about making that shift from the current telco data center to modern data centers where your infrastructure is behind an API where it's fully automated and the kind of data center that you can then build modern applications on. So virtualized applications that scale out, that iterate quickly. But they still need carrier grade, right? They're still special, they still have all these requirements that they think no one else has. But interesting enough, there is real industry consensus right now around building these new data centers with an open-source stack. So if you think about some of the pieces that these data centers have been built, it's open stack, it's things like the SDN controller called Opens A-Lite, Open V-Switch, it's Linux, it's KVM, it's Puppet. It's all of these things that we're building other data centers with. The telcos are also going with these too. And so one of the interesting things that's happened in the last year is Linux Foundation and a bunch of telcos and telco vendors and companies like Red Hat founded this organization called OPNFE. And the idea here is not so much to build a new software but to build kind of a reference architecture for the telco data center using all open-source components and really get collaboration across the telco industry on this. So why are these telcos doing with open-source and why are we seeing open-source just generally dominate on this space right now? It's all about collaboration and co-opetition, right? If you think about the spot the telcos are in right now, they realize that to get out of that and to survive I guess in the long term, they need to collaborate together. They are in competition with each other but they're actually more jointly in competition with a broader change in the world right now. So collaborate together and they can all benefit together. But also I think there's a growing understanding that diversity in projects, diversity in collaboration and diversity of interests really drives some really interesting innovation. If you've kind of got just homogeneous interests in a project, you know, you don't really get new and interesting stuff coming from that as much. But it's also about sustainability too, right? If you think about something like Linux, for example, it's been around a long time and we all know it's gonna be around a long time too. And that's what you're really aiming for. And also a demand on the part of say telcos, if you think about them purchasing for vendors, they really want the ability to not be locked into specific vendors. They want the ability to move between vendors and that's something generally open source delivers on it. So I guess what I'm here to say today is there's a lot going on here. I'd encourage you all to really think about being involved here if you're not already. Choose your mission, play your part. But I guess there's different ways of thinking about how you might approach this, right? You might wanna just get involved and make a bunch of money. And I know that's not the most politic thing to say here, for example. But what I mean here is there's a lot of, there's a lot being invested in this space right now and why not have some open source people and make some money out of it? Some people that actually understand how open source works. But also hopefully have a bit more of a higher purpose here. I guess I'd describe my purpose in this space as, it's about building diverse communities, dedicated to writing and maintaining software that solves problems users care about. There's a lot in that statement, right? There's a lot to unpack there, but hopefully there's a lot of stuff that you get to resonate with you guys there. But I think listening to Eben's talk in particular on Tuesday, it's also about being involved in influencing a future of IT which empowers and enables the future free societies that we all want. So there's a lot going on here. There's a lot relevant to all of us here and I'd encourage you guys to get involved. And thank you. Right, well, we actually are running under time so I'm gonna invite our three speakers back up and we have actually got 10 minutes for questions which is also at you first, wow, that was quick. Hi, it's actually a question for Carol. Seeing a lot of Twitter comments about this one actually. With the summer of code, we've only got about 1% of people in the Southern Hemisphere participating at Google considering setting one up perhaps that the Southern Hemisphere could participate in. We actually run a program that is during the Southern Hemisphere summer right now and it's called Google Code-In and it's for 13 to 17 year olds and it actually, I think, don't quote me on it, just everyone on the live stream just don't quote me on this. I think it either ended this week or last week and so basically we're a team of three and one of us runs Google Summer of Code and one of us runs Google Code-In so unfortunately the answer is we don't really have the time constraints to be able to run yet a third program, not at this moment but we still have lots of participants from the Southern Hemisphere every year in Google Summer of Code and I hope that people are able to make schedule concessions and if they're not then we also have OPW available to them as well to participate in and there's other opportunities too. Okay, my question is for the three of you. First, Carol specifically is to do with the students as well. How many, what sort of diversity do you have in your organizations? And I'm not just talking about women, I'm just talking about people with disabilities and people who are probably not, I would say maybe not as widely accepted in the world but they do some fun things. So like I said in the talk, we're seeing a trend of more and more students from developing nations, underrepresented nations, which is awesome, in terms of women, we have about 12, 13% women who participate in Google Summer of Code. In terms of disabilities, I honestly don't know because we don't ask people to disclose that to us when they participate in the program so I don't really have statistics on that. Sorry. So I'm probably the worst person at Facebook to ask that. We have Australians. No, we... I get constantly reminded by my British coworkers that I'm a convict, so I don't know, is that cool? We have people from all over the world all across Facebook infrastructure. We have a very high percentage of females for an engineering company, like the line, the group that I have here is all of Europe and America and myself, so I guess we're kind of diverse but I can't give you numbers or anything because I don't know it. Oh, I actually have a mic here. Yeah, in the open-source world generally, I think we're doing pretty terrible. And at Red Hat, we're pure play open-source companies so it's pretty reflected within our organization. You know, the way I like to think about it and talk about it is, you know, within our communities, and Leslie gave a great talk yesterday and has given it before about, you know, taking about our privilege and unpacking our privilege, you know, it's a really important thing to think about within open-source communities and, you know, I like to bring it to, you know, very easy to understand examples, like just simply if you don't, say, speak English or if you're, you know, from a culture that you maybe perhaps not want to bring yourself forward so much, it's really, really, really difficult to get involved in some of our communities and influence those communities and anything we can do to, you know, to bridge that gap and be more welcoming communities and more understanding communities. I think that's great, but we have a ton of work to do. Can I just add a little something from our diversity program from this year as well, just to give you guys an idea of what's been going on? You people, you fellow humans. We funded, it's, you know, you guys know, I work for Red Hat, but within the conference I tried to step away from that, particularly when we've been doing selecting of the program, the keynotes, all the bits of the process, I have to be objective. Even within the diversity program, it wasn't until after we funded, people, we realized we'd actually brought in two female interns from Red Hat, one based in India and one based in Dallas, Texas. That's an example of what we're doing in our intern program. We also ended up bringing in an open-source engineer of VMware based out of California. So this is awesome to see these businesses investing in intern programs and in open-source programs in a diverse manner. So that's really, really good. And again, I wanna thank the people who helped fund that diversity program for us. Right, any more questions? Gentlemen from Facebook, I just wanna know whether Facebook, I just wanna know if Facebook's putting any pressure on hardware vendors, lower on the stack to open-source, and I go with open-source harder, open the firmware for the biases, open the whole stack from top to bottom? We put a lot of pressure on hardware vendors. We've suggested that many times, and we've seen some response. So Juniper is one example. They now have an open compute switch that you can purchase and put Junos on it. I don't think Junos has been open-source, but they've made it support on the open hardware. So that's one part, one example I know. The other vendors I don't know yet, but we have suggested we hope by us doing this and more white box switching becoming more pervasive, it will happen. You're all hung over this morning. What's going on, come on. Not the question. Yeah, it's all hung over. This one's primarily for Facebook and for Google. I'm reading a lot about companies complaining about shortages of potential employees and the like. Red Hat, I've noticed, have quite a good remote working contingent, but companies like Facebook and Google want you to move to Sydney in places to work. I can answer that very quickly, which is I have no idea about any of the hiring that happens at Google. I focus entirely on Google Summer of Code, so I don't have an answer for you, I'm sorry. So Facebook, we have our locations. Sometimes people don't wanna go to those locations to work, just so I've understood it correctly. You think Facebook should open more locations to get more people or like? I think there's a whole world of talent and they shouldn't all have to move just to work for these places. There are a lot of... I agree, but I'm not an official like at that level at Facebook. I chose to move to Menlo Park or in San Francisco in the United States of America, got there. But it is an American company, it was started there, so it is American-centric. Google's growing pretty well. I don't know what our plans are on that, so I don't know how to really answer your question, but yeah, I agree with you. There is a world of talent. I know trying to recruit Australian friends and people down here in the Southern Hemisphere, not everyone wants to move to America, but it's personal choice, I guess. I quite like living in America, it's good fun. I will probably come back to Australia one day, though. The way I look at it is I work from Dublin, Ireland, I've been working from my home for the last 11 years, so with our open source projects, to be a part of these projects and make an impact and make an influence, you can be anywhere in the world, and that's valued at a lot of organizations. So that's where we go at right now. We value people around the world who work remotely and it's a model that works really well and very compatible with working on open source, I think. I was just gonna throw in, I'm not in an official capacity, but we have a very strong engineering office in Sydney, Australia. We have, this is Google, and we, sorry, Google has a very strong engineering office in Sydney and we are hiring people all the time. Similarly in Europe, we have quite large offices, so it's true that there is the largest body of engineers in the US, but that's not true, that they're the only ones. Right, this isn't turning into a hiring fast, but if you wanna go and talk to them about a job later, that's up to you, and I think Carol's got some things to give away at Red Joe later if you wanna go and find her. Now, we've got one last question. So this is less a question and more of a call to action. With the comment that maybe there should be a summer of code for the Southern Hemisphere, and obviously it doesn't look like that's happening anytime soon, but we have representatives of many companies here, some of whom have fairly large presences in Southern Hemisphere countries. So maybe someone else could step up to the plate and run a summer of code for the Southern Hemisphere. I would love that, please, let's do that. I've got a full-time job, surely somebody else could take it. Catalyst locally in New Zealand, actually run a whole series of programs through our summer, particularly our student program. So there's things going on, we need more of them, but it's awesome we've got one of our partners is actually running a local program. Yeah, I don't know if it's obvious or not, but the Google Summer of Code model is open source, and so anyone can replicate it, and we would welcome you too, and people have, and it's freely available for anyone to model, however they would like. And oh, by the way, the software that the program has run on is also open source, so if you need program administration tools, you can use that as well. Okay, I want you all to thank our three participants this morning. I hope you got something out of it. I know what I did, it was great. Enjoy your morning, morning tea is being served. You wanna come and find Carol? Hopefully she'll have a few pens and giveaways during the day. Don't bug her too much. Thank you.