A look at the tools and practices used at Facebook to support the #2 site in the world.
The challenges of running a website are different to managing downloadable software or an open source project, and require different approaches to version control. This session will take a look at the counterintuitive idea of shipping trunk and branching in code, which allows your developers to rapidly iterate in production without upsetting your operations team.
Apache Traffic Server is an Open Source project implementing a caching HTTP proxy server, donated to the Apache Foundation by Yahoo! We will examine the technical details behind TS, and three ways TS is used to speed pageload times for a very large website.
The Meebo Bar is served to more than a hundred million users on thousands of websites. If it performs badly those websites suffer. It *has* to be fast.
In "Building Fast Webapps, Fast" we will tell stories of dealing with performance challenges while building the Meebo Bar. Listen in and learn how you can take ownership of your web product's performance without neglecting feature development.
In 2009, Yahoo! introduced an overhaul of its flagship site, www.yahoo.com. The move saw a shift from a mostly-static page to a dynamic, customizable, highly interactive page. What didn't change? Perceived performance by users. Learn the techniques that allowed a smooth transition for 100 million monthly users.
Change management is the combination of process and tools by which changes are made to production systems. Approaches range from cowboy style, making changes to the live site, to complex rituals with secret incantations, coming full circle to continuous deployment. This presentation will highlight milestone practices along this spectrum, establishing a matrix for evaluating deployment process.
Find out what the Chrome team is working on to make the world's fastest browser even faster.
In this session you'll learn about measuring performance of cloud architectures. What simple off-the-shelf software is available? What performance indicators should you look at? What do those indicators tell you about cloud-based and traditional components? What kind of performance can you expect out of Amazon Web Services? Learn to measure, understand, and act on performance indicators.
Many problems in web performance and operations are deeply cultural and only tangentially technical. Sadly, engineers are often very bad at creating cultural change. This talk presents four cases where clever people have created cultural change, and draws out the common lessons contained therein.
This talk inventories high-scale service infrastructure costs and some of the innovations driven by optimizing for work done by joule and work done per dollar.
Users have high expectations of what an online video experience should be that current technologies cannot always meet. This talk explains how to choose the right platform to deliver video at scale and ensure the best possible user experience.
Display advertisements often violate the criteria for optimal web page
performance. This discussion will portray why this is often the case by
taking a look at the array of entities that can be involved with the
creation, delivery and tracking of display ads.
Ads, widgets and other third-party content bring many benefits to your web pages and users. However, they often slow down your pages. We’ll share data on how page speed is affected by such content. We’ll also discuss recent work at Google to make ads as fast as possible, and what site owners and third-party content providers can do to make sure pages are not slowed down by them.
Drizzle is MySQL redesigned to work in the modern world. One of the most important parts of that is designing for huge deployments and all of the operational and logistical challenges associated with that. Come and hear what we're doing to make your life easier!
Browserscope is a crowd-driven test framework for profiling browsers. The goal is to track browsers becoming faster, safer and more consistent for both users and developers. This talk will dive into some of the progress we've seen as well as how you can use, and others are using, Browserscope to store and present their test data pivoting on the user agent.
Google web search sees a lot of requests from modern browsers with a missing or mangled Accept-Encoding header. Google now tests these browsers' ability to understand compressed content, and if successful, sends gzipped responses back, resulting in a better user experience.
Keynote’s demos at Velocity demonstrate how you can improve your end-user’s experience for the next generation of Internet content, applications and services – the “Next Web”.
Much of the FOSS stack used for running webapps e.g., memcached, mysql, rails, has scalability limitations that are masked by the widespread adoption of horizontal scaleout. But as traffic growth forces the need for cheaper multicore servers, multithreading scalability becomes a real gotcha. This talk will show you how to see such issues ahead of time, quantify them, and improve on them.
I have made a map of everything involved in getting from your computer, via your ISP, to a web site and back to your eye balls. It's exhaustive, but that's the point. Where the heck should we optimize? This talk explores the important features of the web and where we should optimizes. Learn about the different ways people are optimizing the different pieces of the puzzle.
Twitter has grown from a micro-blogging service to a real time communications system that is an underlying part of the web. With the rise of the service has come a scaling challenge, one that the operations and engineering team has met year after year. In this session, John will share some of their challenges from this year and to revisit the best practices that they've discovered along the way.
Infrastructure is code – the separation between how you manage your infrastructure and how you build your applications is disappearing. Adam Jacob, CTO of Opscode and primary author of Chef, will teach you what this means in practice – through showing how to deploy real-world applications with Chef on EC2.
Cheryl will review two case studies where Yahoo! was able to solve problems of scale efficiently by using tools developed by Yahoo! engineers.
commercial websites and compare this behavior with the benchmarks.
Featured by Steve Souders at Velocity Fall 2009, Show Slow is an open source tool that helps keeping track of performance metrics gathered over time using Yahoo’s YSlow and Google's Page Speed Firefox extensions using web-based *ESTful beacons. It provides easy to understand graphical representation of metric changes over time that makes decision making quick and easy.
Money can buy more bandwidth, but to improve latency, you need to increase the speed of light (or electricity). Latency is a problem that all network apps face, but web app designers need to worry a little more. Find out why and what you can do about it.
Demos from dynaTrace, Firebug, YSlow, and Page Speed.
Demos from HttpWatch, AOL Pagetest, Speed Tracer, and Fiddler.
This session will help you build a complete web monitoring strategy. We'll cover the many different metrics you can collect, from latency and uptime to usability and navigation - and show you how to tie them to the goals of your web business.
Ads are painful to deal with, but they pay the bills. Ismail Elshareef will share the lessons he learned on how to lessen the impact of third-party components, including ads, on site's performance. At Edmunds, they have iterated over multiple solutions until they found an optimal one that works and they'd like to share that with you.
For the last five years, we´ve talk, discuss and test all the web techniques for high performance. What about mobile web? How mobile web browsers do request and page rendering? How to optimize to the maximum the mobile web experience?
Last year we made Facebook twice as fast, but it turns out that that's the easy part. The hard part is keeping it fast as things are constantly changing. I'll be talking about the code frameworks, tools, and engineering culture that come together to keep our site moving fast as our engineering team moves fast to build new products.
Tim O'Reilly shares his insights into the world of emerging technology, presenting his take on what matters most - and what will be most disruptive - to the tech community.
Last year at Velocity, Hooman Beheshti presented the findings from phase one of Strangeloop’s long-term research into the relationship between web performance and business benefits. The results were also published in Watching Websites. Since then, we’ve received a barrage of questions from the web performance community, which fueled phase two of our study.
Many developers wrongly assume Google.com- and Amazon.com-like performance and consistency when they enlist cloud computing services from these and other major cloud providers. But the truth is, developers must do the due diligence and insist on performance guarantees from the cloud provider.
More and more webapps are storing sensitive data for their users. Personal finance company Wesabe has open sourced a piece of software, Grendel (http://github.com/wesabe/grendel), that we think can help many sites protect users’ data in a simple way. This talk will cover the ideas behind Grendel, how and where to use it, and its limitations and future directions.
Running PageSpeed - all checks. YSlow - A. So what's next to do? Well, you can cheat! You know that time is relative - one hour with your mother-in-law feels much longer than one hour on the beach. The same way you can bend the perception of the page loading time and make the page feel faster, even though the RTT is the same.
The story of how Picnik removed a single point of failure by doubling it's operations staff from one person to two. We'll talk about when we made the move and how it worked out.
Internet traffic spikes aren't what they used to be. It is now evident that even the smallest sites can suffer the attention of the global audience. This presentation dives into techniques to avoid collapse under dire circumstances. Looking at some real traffic spikes, we'll pinpoint what part of the architecture is crumbling under the load; then, walk though stop-gaps and complete solutions.
Different organizations mean different things when they talk about scaling. Sarah will offer some tips about a few different ways that this term is thrown around for MySQL databases. Each different dimension - data volume, read volume, and write volume - present different challenges to the operations and development staff working with the system.
It is possible that, working together, browsing the Web can be as fast as turning the pages in a book. This session describes initiatives for achieving this including faster browsers and improvements in TCP, DNS, SSL, and HTTP.
Network latency, not bandwidth, dominates to-the-glass latency of many web applications. This talk explores the history and dark corners of TCP to examine how decisions made in the 1980s effect the speed of your applications today.
By now, we’ve all internalized Steve Souders’ rules for optimizing web performance, but the question is: do you need to spend 6 months and raise an army of top developers to make your sites fast by default? In this workshop, we’ll subject an unsuspecting website to real-time optimization, following Google and Yahoo’s rules for high-performance websites.
Mobile web users are demanding. More than half say that they expect sites to download as quickly on their mobile devices as they do on their home computers. Three out of 5 say that poor performance will make them less likely to return to the site. And 40% say they'd likely visit a competitor's site next.
Your website has out-of-control CSS bloat. You know your performance is being impacted, but how do you move from organic CSS with no particular architecture to something lighter, more logical, and easier to maintain? In this session, Nicole Sullivan and Stoyan Stefanov will show you how they improved the CSS at Facebook and Yahoo! Search.
The only sure thing about your site is that it will someday go down. More importantly, thanks to Twitter the news of that downtime will not stay a secret for long. What matters now is how you respond to that downtime. A little preparation up front, a focus on transparency, and authentic communication are the keys to surviving downtime with your reputation intact.
Shopzilla is one of the largest and most comprehensive comparison shopping networks on the planet. Delivering content to millions of users, thousands of times per second – FAST – is not just good for business, it's a competitive advantage. This talk will provide updated metrics on the value of performance for Shopzilla and our techniques for achieving, measuring and defaulting to high performance.
Jesse Robbins and Steve Souders welcome you to Velocity Conference 2010.
Jesse Robbins and Steve Souders welcome you to day two of Velocity Conference 2010.
At this presentation, Hotmail will share with the Velocity Community its key findings surveying last-mile bandwidths and network latencies around the world.