 We're back here live, this is SiliconANGLES, theCUBE Covering Velocity Conference. We're here with the co-chair, Steve Souders from Google. High Performance Engineer, welcome to theCUBE. Thank you very much, glad to be here. Excited to be here. I know, you're looking forward to it all week, the keynotes and all your appearances. You must be super busy. Yeah, very busy, but you know, it's only three days. And so, you know, I can maintain that level. Tomorrow I'll be a little tired, but it's exciting, high level of enthusiasm here. And that's something that keeps me going, is people are always excited to talk more about performance in the companies here. We've been doing the commentary here in theCUBE all week, and it's funny, we're trying to put our finger on, what kind of conference this is. We kind of know the conference. Great community of people. It's a kernel of some really tight developers. It's certainly a no BS zone in terms of that's concern. And, but you know, DevOps is cloud. Performance is performance. Monitoring, setting all the innovations there. And at the edge, you got user experience, design, and performance, which is browser on one hand, mobile on the other, but it's a holistic integrated world now. So DevOps drives this. So the trade-off between design, stuff that you mentioned on your keynote, that's the theme here, but it's not like a cloud conference, and that's not for service. It's not, oh, how to do great UI. It's a little bit of both. Was that part of the design, or is that just demand, or is that, how did you look at that? It really came from the people who first sat down with Tim O'Reilly and pitched the conference, is these were just things that we cared very strongly about, and we knew we were connected with other people that had a lot of big companies that had that same passion for this, and sharing best practices and lessons learned. And we just knew that there were lots of, the people at every company who thought that way and could benefit from these insights. And so it does kind of transcend across a lot of different spaces, a lot of different types of roles at companies. So you've had the job as Chief Performance Dude at Yahoo for seven years, and now Google. Two different kinds of performance objectives. One's get the webpage to load fast, the homepage on Yahoo, and Google is just the faster it loads, the more quicker people get to their destination, direct navigation, and ad clicking. So it's just the numbers came, right? The faster it can get there. But now it's beyond that now, right? So you've got mobile now, integrated, you got Gmail, I got other services from Google, and people have kind of an integrated mindset. So what is the new, if you could put your finger on it, what is the new performance variables that people need to be thinking about right now? If you're looking at the holistic DevOps to design? Yeah, I think the big thing that everyone should be thinking about now is mobile, right? And it's not just mobile in a year or two, it's going to be TV screens, playstations, player screens, maybe TV, certainly mobile and tablets falls in there. And in one sense, it's funny because people will kind of, people are totally right and they'll say one thing and the opposite thing and both of them are true. And that is they'll say, isn't the best practices that we should do on mobile the same as desktop? And there's a lot of overlap, there's a lot of truth to that. And they'll say, isn't what we need to do on mobile very different than what we do on desktop? And there's a lot of truth to that as well. And so what I tell people is we're still, even though we've been doing mobile for years now, we still haven't figured out all of the variables that go into a good mobile experience, good being fast and reliable. And so we're still learning our way and that's why it's great to have conferences like Velocity where people who are out there building things, they're being used every day on big websites are here sharing their best practices. Yeah, I want to give you some props because when I was truly trying to did my homework coming into Velocity, I'm like, I love this show. How do I tell my tech geek friends who aren't in the weeds, who aren't under the hood, all the geeks here or the tech athletes that are here, what this is. And it's a confluence of many things. And even analytics, there's a little bit of big data here, right? I mean, the treasure trove of data that's now available to be integrated into not only understanding design and criteria, but also what's working, what's not working, what's the feature, what's not the feature, that's a UI issue, that's a design issue, that's a DevOps issue. So again, it's a confluence. But the real exciting thing now what I want to say is, is that this whole HTTP Archive initiative that you put together is really fascinating because there is a treasure trove of information out there. Could you share with the folks out there because I'm looking at the data on our dashboards and the HTTP Archive blog post that you wrote or has been written about is the most popular shared item on Twitter for your conference. Well, it's behind SiliconANGLE TV but we don't count that, that's our site. But in terms of the geeks, they want to know about this. There's a little data science involved, there's a little bit of crowdsourcing, can be a little Kaggle competition down the road. Why did you do that project? It's certainly well received. We talked about it earlier today with another guest. Tell us about the HTTP Archive. Yeah, so I started that about three years ago and it came from a very simple motivation. I'd been trying to get the project off the ground for a few years, is when I would talk to various development teams, I just found that no one was tracking the metrics that were really important for their website's performance. And so I would go back to teams every week or every month and we would do a check-in on performance and how it was progressing. And they would say, yeah, it's gotten worse. We don't know why. I would say things like, well, have you added a lot more domains? Have you added a lot more JavaScript? And they would say, well, I don't know. And these were people working on huge websites, like the top websites in the world. And they didn't know how much JavaScript was in each page that they were building and pushing out to the whole world. And I said, we need a tool, we need an easy way for people to track these kinds of stats across all the world's popular websites. And we could build something like that and have each company decide whether or not they want to adopt it. Or I took the other route, which is, let's build this thing and let's just run it on the entire web. Now I can't run it on the entire web. It's a nonprofit project. It's now part of the internet archive. But we do, right now we're up to the top 300,000 URLs worldwide. Our goal is to hit a million this year. And so you can go and if you're in the top 300,000 websites, you can go and we probably have stats on your website. Like, do you use the expires header? How many requests do you use? How many domains do you use? How much JavaScript is on your site for the last two and a half or three years? And if you're not in the top 300,000, there's also a feature where you can add your website to the list of sites analyzed. It's just a brilliant idea. It's almost intoxicating to kind of riff on this because it's almost like the caveman kind of walking a wreck over the years. It's like, you can actually look at, you know what, you were once used JavaScript this way and then you can see efficiencies, right? So you can see the evolution. But I think more importantly, I think what's brilliant is that search was genius in the sense that I'm scraping metadata off the page content and surfacing that up for discovery, user experience. And then, but what you do with this that I find certainly fascinating is that you're basically taking the metadata of page structure, not the content, and surfacing that as a big data opportunity to create automation around analytics to see what's good. So I mean, you can almost write a search engine just for that. Is there a search for that? Can you give me the best page designs? Or can I query? No, we don't have a search around that. The other thing it tells you, besides just the data, the analytics that it generates is it also helps identify trends. And that's really important for the people who are in the audience here attending this conference. What kind of trends? So we had a great talk about that. The guys from Brightcove, which do VideoJS. So they're a video development company. And they were talking about how the HEP archive is tracking the adoption of HTML5 video formats as well as Flash and how Flash is dropping and that HTML5 video is actually going to be supported and more popular by more users than Flash probably this year. And so the takeaway for that kind of trend, companies should be moving to those newer formats. Developers should be coming familiar with those new programming techniques. It's a sentiment engine. It's an investment opportunity for them saying, hey, do I want to spend resources building out this technology or contributing to this open source project? It truly is a democratization. I mean, that's data driven. I mean, I just want to say, I really liked that project that I think you hit a home run with that. And then I wanted to do a shout out. You mentioned the popular blog post today. That was actually from my office mate, Ilya Grigorik, where he took the data in this, because we're a nonprofit part of the internet archive, all of the data, all the code is open source. He took the data and loaded it up inside of Google's BigQuery engine. And so I've got this MySQL database that has all this data and it takes me minutes to do a query. And now he's got all the data available to anyone so you don't have to download it and you can run a query. What takes me two minutes, it takes two seconds. We interviewed Ilya earlier on Grigorik. He was on theCUBE earlier. He and I had a fascinating conversation about that. Again, it's a search engine. It's BI for page construction. So is that going to help us in the future with design? Yeah, we're going to see things. I mean, another big part of design and especially the perception of speed and design is it affects the user experience is images. And so we're going to see a lot of, we're seeing a lot of trends in image formats. We're doing more research on popular image formats. And so Yoav Weiss is another guy who's using a lot of the data out of the HTTP archive. And he's doing studies like how many of the images in the popular websites are progressive JPEG format, which is good for some performance benefits. How many of them are sized appropriately on mobile devices? So yes, we can get a lot of great trends for design and user experience there too. I want to ask you, Steve, as we get down to the wire here, I know you got to go, you're really busy and we're going to be booted out at four o'clock today, hour ahead of our planned schedule. So we're trying to get all the content out there, all the data out of your head and into the audience's hands on Twitter. What is the biggest surprise here at the show that surprised you? Is it good or bad that you didn't think was going to pop out of the woodwork? Is there anything that you can share that folks anecdotally, personally, or content-wise? I think just one serendipitous thing was that we have two co-chairs, John and I, and we independently found keynoters, both of them from the Obama campaign. So I thought that was funny. We actually sat down to do scheduling and we knew before that, but we said, all right, well, let's make sure, should we put the two Obama guys on the same day for opposite days? That was one thing. But I think the main thing, I shouldn't be surprised about it anymore, but I think one thing is just to come here and see how large this community is, we're sitting here at one end of the exhibit hall. We can't see the other end. There's so many people and so many great companies here. So I think, you mentioned before, it's hard to pigeonhole this conference and it does, it really does touch on so many different areas and transcend so many different roles at companies and yet it's really popular that the speakers and the companies are awesome. Well, I would say, I think you guys have a good form. I wouldn't be discouraged by that. I think that matches what we're seeing in the marketplace. You're seeing, even on the IT side, which is a little now just getting in the bandwagon, to the progressive early adopters who are doing the pioneering, it's an integrated world. You look at DevOps, what DevOps has done. You're looking at what's going on in the UI. The consumerization of our world, consumerization of IT, this is now the new normal, the integration, the collaboration. So I think you got a good form here. I think it's going to break out. It's going to be really, really big. And obviously the browser stuff has been phenomenal. Obviously Chrome is amazing and Firefox announcing their native support. WebSockets is now really important. Node.js. I mean, these are advancements that are just free. WebRTC. WebRTC. Again, who would have thought WebSockets would be that killer now? WebRTC. This is just the stuff that's been around for a while. But now, as you get more native into the browser, good stuff's going to happen. Final question, what should people walk away with this year from velocity that you want to share with them? I think the main thing is that performance does have a dramatic impact on the bottom line of whatever business you're in, whether it's something like impressions for a search engine or downloads for a free software product like Mozilla or Revenue, if you're in eBay or in Amazon, those companies are here as well. And so performance does really have an impact on that. And I think the other thing is, I think we're going to see more of a focus over the next year on something really, really interesting to me. Coming from a coding and number crunching background is how we need to start getting a better handle on tracking the user's, the user experience and the user's perception of a website and how the user engages with a website. So there's a lot of challenges in that. It brings in, you know, psychologies, you know, psychological factors, but also we really do like to quantify performance if we're going to try to optimize it. And how do you quantify some of those intangibles? You know, we started SiliconANGLE blog four years ago. You know, I decided I didn't want to have any ads. It was completely data driven. Now theCUBE is all open source content. We don't charge for anything. But we have a nice data practice. But I think our tagline from the beginning was computer science meets social science. And I think what you just talked about was there's a social aspect here. So a societal benefit. And Kate Matz, I was on earlier, she talked about the pop forms. Really interesting angle on the leadership issues. Just the human resource opportunity for not just job training. That's kind of obvious, right? People can, you know, in cancer. But how to manage in this environment? Because it's innovative and change is here. So congratulations, Steve. I really appreciate you coming on theCUBE. My pleasure. This is theCUBE, siliconangles.com, the exclusive coverage of O'Reilly Media's Velocity Conference. Steve is the co-chair with his partner, John. Great program. We'll be back with our next guest after this short break.