That black box of HTML <video>, appealing in its ability to offer native browser video playback, hides many performance dangers if poorly implemented. This session examines optimization strategies for single bitrate video and then looks at techniques and best practices for delivering adaptive multi-bitrate content through the new MediaSource Extensions (MSE).
Etsy published its first quarterly Site Performance Report in August of 2011, and with each post we have improved our tools, methodologies, and best of all, the site's performance. This talk will combine helpful tips and lessons learned to make a case for sharing your performance data with all of your stakeholders.
In web operations, monitoring is the main interface to our applications. A monitoring system alerts us, interrupts us, awakes us. When it does, maybe the site is on fire (signal), maybe a new application broke old monitoring assumptions (noise). Noisy monitoring leads to outages, frustration and resentment. How do we cut the noise and turn monitoring from necessary evil into operational strength?
This lightning demo will show you how to complement your favorite page performance tools like Page Speed, YSlow and Web Page Test with Keynote KITE by analyzing dynamic transactions spanning multiple pages and user interactions. We’ll also show how Keynote MITE can help you understand mobile web page performance on virtually any mobile device—including tablets.
The web is ever changing… browsers are evolving, new protocols are emerging and mobile continues its relentless rise. We’re already starting to bend some of the original performance rules and as the web changes further will our current good practices last, or will some become barriers that hinder performance?
Most of us think of Jenkins as a build server, but it can be so much more. The flexibility of Jenkins makes it an interesting tool to add to your operations toolkit.
Make Jenkins the next member of your Operations team, automating tasks like syntax checking before restarting services like Nagios and BIND, ensuring that tests pass before affecting production.
Dropbox is used by over 200 million people to store their valuable data. Achieving such scale inevitably comes with unexpected challenges, tough decisions, and valuable lessons. In this talk, we'll discuss why and how we started by being hosted on AWS and colocation services to our current hybrid model that combines the unique strengths of AWS and dedicated hardware.
When raw data becomes overwhelming, we turn to abstraction to understand our world. In our systems, the data is always overwhelming. Solutions like summary statistics have come to our rescue, and they are good -- up to a point. In order to truly understand our systems, we need to know when and how to sidestep those abstractions, to get deep, detailed performance insight.
Stack Overflow is a developer-centered Q&A site handling 3 billion requests per month using a variety of platforms and products to deliver content. With a small sysadmin team, building our deployments, applications and architecture for ease of management and maintenance is key. Most of what’s used is an application of widely available open-source, or written in-house and open-sourced for you.
Program Chairs, John Allspaw and Steve Souders round up the conference with a recap.
Oftentimes security seems like a daunting challenge which many organizations aren't sure how to get started on. However, the engineering principles inherent to devops and continuous deployment lay the groundwork for a number effective security mechanisms.
Synthetic performance monitoring is a key piece of our monitoring toolkit, since it provides stable data over time and more in-depth information than you can get from RUM. Everyone loves WebPagetest, but the public instance isn’t great for high volume testing and trending over time. We’ll explain how Etsy uses a private instance to fill this need, and share the code that drives our tests.
In this session, learn how to measure and monitor a content delivery network that's operating at a very large scale. We’ll review best practices on how to measure and eliminate bottlenecks to significantly improve global purge-from-edge time, track outliers during core cache/proxy engine deployments, and identify and find our way around quirks in the TCP stack.
Keynote by Fred Wilson, venture capitalist, Union Square Ventures.
This session will cover the evolution of the technologies, tools, and team that power the Evernote service. It will include some of the joys (and woes) of dealing with with a diverse set of client applications.
In Jan 2012, HTTP Archive reported Intuit.com as taking 15s to load. By Apr 2013, this page was loading in 3.6s, with 5 additional sites showing significant improvements from the previous year, some as fast as 2s. How did we do it? What worked or didn't work? Come learn about what we did, the techniques we used, and how faster pages impacted our business results.
Web performance is usually measured until the load event fires. That metric is diverging further from the user experience and is becoming less useful. We will explore three different experimental techniques for a better "user experience" metric.
Examine Google’s vision of the cloud platform and take an inside look into Google Compute Engine, which allows you to run large scale computing workloads using Linux Virtual Machines on Google’s infrastructure, cloud basing your infrastructure.
Braintree is a payment gateway, so downtime directly costs both us and our merchants money. Therefore, high availability is extremely important to us. This talk will cover how we do HA at Braintree on our Ruby on Rails application, including no-downtime deploys and load balancing across services and ISPs.
The real challenge of running an effective post-mortem is the complex, messy, all-too-human *feelings* of everyone involved. This talk gleefully tackles that challenge head on, and lays out effective techniques for helping a team to learn as much as possible after a public (and generally, pretty shameful) failure.
In the startup world, the most pressing issue usually isn't, "how to we prepare for a once-in-a-lifetime storm?" Risk management is always stressed at Velocity, but businesses don't always buy into dedicating resources to low probability events. Let our comedy of errors during Hurricane Sandy help to convince your bosses that infrastructure and redundancy are actually worthwhile investments.
An Informative Workspace can benefit almost any team, but it is sometimes difficult to allocate the time or budget to set it up. This presentation will describe the steps needed to easily create an inexpensive, state-of-the-art informative workspace using a Raspberry Pi, inexpensive TV, and a few basic electronics.
Optimizing mobile applications in real user conditions and on a wide range of devices is imperative for any company devoted to quality and performance. In this session, we’ll introduce you to Appurify’s platform for mobile automation, which significantly speeds the process of testing, debugging and optimizing applications. Come learn about one of the newest development platforms for mobile!
For many high-performance web applications, badly managed memory can create a slew of performance problems, especially for mobile! Join Colt for a quick lightning demo where he'll be showing off some very quick, powerful techniques in Chrome DevTools to help you track down and fix your memory problems.
In this demo, Guy will demonstrate the integration between WebPageTest and the Appurify platform. With this integration, you'll be able to easily run mobile web performance tests straight from WebPageTest without buying a single device, by leveraging short term, on-demand access to a wide set of Appurify-managed phones.
A demonstration of how to use the HttpWatch mobile app to record mobile performance data.
At Qualcomm we decided to create a benchmark tool that should answer exactly that question. This demo will show the benchmark in action, review its components and get a glance into differences in browsers performance that it can uncover.
Pat will be demonstrating some of the latest test comparison features in WebPagetest.
In this talk, we will discuss the strategies we implemented to make it faster for users to accomplish their tasks at Bing. We’ll also go over some novel ways Bing has optimized the time it takes to load our web pages.
You can obtain big page speedups, particularly on mobile, if you re-structure your pages to prioritize the critical CSS and images and lazily load the rest. We’ve been studying the impact of these changes on large numbers of sites; in this talk we’ll present what we learned.
This talk will introduce resilience and a model of system dynamics useful in analyzing failed and successful event management and offer an explanation for why our systems run at the edge of failure.
Over 90% of responsive designs are bloated messes that are the same size and some times even larger when viewed on small screens. The main reason for the bloat is inappropriately sized images. In this session, we'll look at responsive images and how to find a comprehensive solution for your organization.
In this session, we'll look at the current state of performance for responsive sites and some of the challenges unique to responsive design. Then we'll look at the tools and techniques we can use to start building faster responsive sites today.
This talk looks at how Facebook has redesigned its configuration management system to handle a massive, dynamic, heterogeneous environment with a tiny team and open source software. We will look at the philosophy we use to manage our systems, the implementation of that philosophy, and how you can apply these ideas to any size server footprint, from a handful of servers to a global environment.
Paperless Post recently replaced Nagios with Sensu, a new and awesome free monitoring and metrics router that is designed with configuration management and cloud deployments in mind. In this presentation we’ll take an in-depth look into why we chose Sensu, how to deploy it, how to scale it, and how to integrate it with your existing tools.
Mike O'Dea and Bill Zurat will discuss how MLBAM used AWS, automation and other off-the-shelf technologies to deliver a product (Minor League Streaming) that customers would enjoy, and was consistent with how MLBAM delivered media in their MLB products.
In this talk, Philip Tellis discusses insights gathered from analyzing real user monitoring (RUM) data from global users with different network characteristics across billions of page views & from different types of websites.
In this talk we’ll go through a systematic approach to optimizing your stylesheets, cover the basics of CSS optimization and show how working with preprocessors such as SCSS can make it easier to optimize your stylesheets.
We will do a hands-on example of using synthetic testing, building Real-User Measurement, Back-end performance tracking and pulling it all together with off-the-shelf tooling and reporting.
The mission of the Web Performance Working Group is to create standards
to enhance aspects of Web application performance of user agent features
and APIs. Some of them are already deployed in today's major browsers,
like Navigation Timing or Page Visibility. This update presents the
latest proposals on improving Web performance.
Keynote by Courtney Nash, Editor at O'Reilly Media.
In the field of operations, whether they be web, cloud, systems, network, or IT, we are finding that there is a problem finding qualified people to help us do more things. Once we've grown old, who will take over for us?