Images matter A LOT for web performance, and the variety of formats, quality levels and screen sizes complicate optimization. Retina displays, RWD and stubborn designers further fuel this fire.
In this tutorial we’ll review tools and techniques that can help keep images from slowing down your page, share data about the results you can expect, and discuss what’s next for image optimization.
Come find out (and see demos) of how to bridge the interaction performance gaps that often exist between native and web apps and create mobile web experiences that delight your users without resorting to heavy weight mobile frameworks.
The Financial Times web app was one of the first, and is still one of the best, examples of how HTML5 can genuinely rival native technologies in performance and user experience. But combining multi-platform interoperability with a responsive user experience remains tough. This talk will cover many of the techniques developed at the FT to meet that challenge and nurture the web platform.
We all know mobile performance is important. But how does it affect user engagement? We conducted a set of experiments with the Facebook app (both Android and iOS) and the Facebook mobile site to see how users would react to changes in page load time, scroll delay, or reliability.
Mobile networks are slow, unreliable, and have high variability both in latency and bandwidth - a nightmare to optimize for! Right? Well, almost. It’s not nearly as bad as it seems once you understand the limitations of the device and the radio network. We're going to build a 4G network, from the ground up, piece by piece, and talk about the performance implications of each and every component.
The mobile web is filled with sites that just feel slow. This talk will help make sure your site isn’t one of them. This talk will cover how to diagnose and solve runtime performance problems on mobile, covering both focusing on optimizing for interaction and data transfer.
Mobile add new challenges for web performance optimization. The biggest problem is usually underestimating mobile platforms and the Facebook HTML5 app gave us some learnings. In this tutorial we will cover the mobile ecosystem for HTML5 websites and apps, how to measure performance on these platforms and how Responsive Web Design may help or harm performance, including live optimization examples.
The CBC serves optimized content to millions of Canadians. We'll share our experience and knowledge of optimizing content delivery for a high-scale and unpredictable audience. We will explain our performance stack from server-side optimization tricks to automated performance tools during deployment. We will discuss our challenges, findings and learnings of continually improving site delivery.
A look into how we tackle some aspects of capacity planning at Twitter.
Common patterns, production examples and operational learnings from how Netflix incorporates fault and latency tolerance into its distributed systems using circuit breakers, bulkheads and other patterns embodied in the opensource Hystrix library (https://github.com/Netflix/Hystrix) and operates them using realtime metrics and data visualization tools.
In the third edition (2012) of High Performance MySQL, MHA usage is described as: "anyone other than Yoshinori who is using it in production, and we haven't used it ourselves." Little known to the authors, we've spent quite a lot of time working on MHA and figuring out its kinks. Today MHA is in use in places like DeNA, Facebook, SK Telecom, and DAPA.
An introduction to Etsy's real time anomaly detection and metrics correlation systems - the newest front-line Ops tools from the team that brought you StatsD.
Want to learn how facebook scales their load balancing infrastructure to support more than a billion users? We will be revealing the technologies and methods we use to route and balance Facebook's traffic. This talk will focus on Facebook's DNS load balancer and software load balancer, and how we use these systems to improve user performance, manage capacity, and increase reliability.
This talk will cover a number of different anti-patterns, from small issues to architectural decisions that will help configuration management users of all levels build more sound configuration management scripts.
You're on board with the awesome. You know you're going to do configuration management and automate All the Things. You might even have picked a tool. Except, now what? Where do you start? The whole world is available to automate and now it all seems kind of complicated.
In this presentation, we will show you how quickly Ansible can be used to set up a PostgreSQL deployment in EC2 using master, streaming replication, backups and monitoring agents rapidly.
As demands for performance increase and the complexity of Internet application architectures grow, operations are abandoned in a perfect storm. The problems are ever harder to diagnose and the need to rapid remediation is more than ever. You need the right techniques and methodologies or you'll be ineffective or, worse, an obstacle. Learn them here.
A walk through for improving overall operational effectiveness by building skills like statistics, software development lifecycle practices, graphing and metrics. This tutorial will focus on skill development and building better products.
One of the monitoring holy grails is knowing whether a system is behaving normally. This is usually considered a hard problem, but it can be done very cheaply at large scale. In this presentation I'll explain my definition of normality, several ways to quantify abnormality, and most importantly, what you can do with the results. There will be math, and I'll publish my slides.
As LinkedIn's traffic grew, it was clear the classical approach to load balancing would not grow with us. In addition to issues with scale, we found many missing key features along the way (cookie routing, DDoS protection, etc). This talk will cover the detail of our migration, things learned along the way, plugins/patches we contributed to ATS, and how other web properties can leverage them.
Deep inside Adobe, the Coretech group builds the enabling tools and shared components that underpin Adobe's award winning products. But what happens when Adobe starts the move away from packaged software and towards Cloud based services? Learn how Coretech became a service provider, delivered a self-service build & deployment platform, and reinvented itself in along the way.
Various performance analysis methodologies for production systems and cloud environments.
chefspec, strainer, fauxhai, test-kitchen, foodcritic, test-kitchen, and jenkins OH MY! The devops toolbelt has expanded and the tools to test are not far behind. Learn how to create a test-driven infrastructure with Chef.
On the Google Consumer Surveys team, 8 minutes after you commit code it's
live in production. This talk will cover our experiences doing this for a
year, what's gone wrong, and how we made it safe using website screenshots
and visual diffs between each release. We'll go over the software that
makes it work, how you can implement the same process, and the positive
effects it has on a team.
The performance of the network underlying our applications can have a
dramatic effect on the experience of our users. This session will cover how to tune some of the interactions between TCP and your application to deliver solid performance over the public internet.
A deep dive into the AWS offerings for running MySQL at scale, with appropriate consideration to availability, failure management, monitoring, trending and configuration management. Real world examples such as Call of Duty Black Ops 2 and Obama for America will be used to highlight key points from a practical perspective.
Johan Bergström, Associate Professor at Lund University, Sweden, will introduce the notion of risk in system design.
We designed the 2012 DevOps Survey to learn the habits and health of the DevOps Community. We received 4,000+ responses, and the analysis yielded astonishing results. DevOps is no longer the domain of large-scale WebOps shops; it is entering mainstream. We will share DevOps practices that correlate with significantly higher business performance, as well as their demographics and psychographics.
How gamedays at the Obama campaign led to an uneventful Election Day.
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.
Velocity is about performance and operations - and the most high performing operations organizations practice both Devops and Continuous Delivery. This talk will show you how to successfully adopt both within your organization, supported by real-world examples of victory and defeat. Revolutionary change requires revolutionary behavior - this talk tells you how to win.
Being great at operations means much more than being technically competent.
To be truly successful you have to understand people. Every one of your teammates and everyone one of your customers is a person. And to get what you want, you have to build trust and meaning with those people.
Firefighting. Not academic where you can debate for a long time. Pressure is constant. Burns many folks out but have to make local optimal decisions and keep pushing to improve.
Operations people frequently communicate and learn through stories and mythology. We're going to examine why these stories and mythology matter, how they shape your culture and why sometimes telling these stories can give your Operation's people nightmares.
Conference talks often cover how to measure, and improve services through dashboards. This talk aims to help the audience use personal metrics gathering and visualizations to reduce interrupts, improve self assessment and manage career growth.
How twitter.com, a high performance web application developed by several engineers, avoids performance regression at every new release? Join us to learn more about how we make it through synthetic benchmarking and RUM.
As the web goes mobile, applications get asynchronous, and functionality moves to the browser, the meaning of “site performance” has moved beyond page load time. We'll talk about picking new relevant metrics, how to adapt existing monitoring tools, and how developers can tune their apps to help track what users are actually experiencing on fully asynchronous pages.
HTTP guru Mark Nottingham from Akamai dives into the details of not only HTTP, but below, into TLS, TCP, IP, DNS and more, with many stops along the way to see what kinds of things can happen to your content as it gets mangled by the Internet.
Nobody likes slow video, but it doesn't have to be that way. In this session we'll explore the most performant technologies and techniques for processing video and delivering it to your viewers, including faster player loading (HTML5 + Flash), bit rate switching, peer-to-peer video, CDNs, fast and scalable transcoding, better video file compression, and accelerated file uploads.
When it comes to analyzing web performance data, which presentation works best? Response time plot? Median value? Standard deviation? Baseline comparison? Histogram? Apdex? With enough data you can get creative with visualizations, but which one brings the most clarity? Come look at real world performance data from many perspectives and find out which works best for your needs.
Monitoring the performance of today's complex web applications over the coarse of the development is far from trivial. This talk discusses the emerging language tools designed for the new breed of web applications, ranging from syntax augmentation/transformation, advanced static and dynamic code analysis, as well as run-time complexity profiling.
Join Sergey Chernyshev, Andy Davies and Rick Viscomi for a hands-on Web
Performance Optimization session. They will review latest performance trends and show how to use
performance analysis and optimization tools to speed up worlds most popular web sites
- bring yours too and they will look at them as well.
SPDY is in the process of becoming HTTP/2.0, but won't be final for a while. This panel of industry experts will offer their perspective on what's been happening as well as what needs to happen after it's done.
While iframes have been the sandboxing method of choice by security folks, they haven't always had a good reputation for web performance. In this talk, we'll look at two techniques that use iframes to boost performance. The first is a method to load third party scripts without blocking the onload event, and the second makes updating far-future expiring content a cinch.
Join us for a snappy series of five-minute talks on the main stage.
Join us for a snappy series of five-minute talks on the main stage.
The Obama engineering team created a new fundraising platform from the ground up with performance as a primary goal. It was 60% faster than the previous platform. When tested against each other the new platform converted 14% better, a more than $34.7 million difference during the life of the platform.
Everyone in the performance community is trying to better understand RUM and how to use it. In this presentation I'll introduce, compare, and contrast the two primary methods of performance monitoring. This will be done by using Synthetic and Real User Measurement side by side on a handful of sites for 3 months and aggregating and presenting the data that the two methods yield.
This presentation will cover DNS performance issues, sharing research and insights on DNS performance and the different ways browsers and devices handle DNS and are impacted from it, as well as tips for better DNS handling that will result with more robust and better performing sites.
We will also cover IPv6, DNSSEC and open resolvers: what is their impact and how to optimize for them as well.
As more companies realize the benefits of real user monitoring, the number of RUM alternatives is growing. Come to this session to get a quick look at leading RUM providers to see their key features and how they compare.
Chrome implements fast, powerful webpage rendering by using the graphics hardware (GPU). But powerful GPU rendering introduces a new suite of issues that web developers need to understand. Attendees will walk away with a clear understanding of the GPU compositor in Chrome and knowledge of how their design choices will impact a page's performance on devices with GPU compositing.
CSS is a simple declarative language. Preprocessors were introduced to overcome it’s many limitations. Developers have long awaited this functionality, and the use of CSS preprocessors like SASS, Less, and Stylus has become ubiquitous on big websites. Yet CSS preprocessors often output some downright ugly code. During this talk, you will learn about preprocessor performance best practices.
We'll look at some of these failures & what they can teach us about how to improve WPO technology and make better use of it.
The W3C Web Performance working group creates standards for improving web application performance including Web Timing, Page Visibility, & High Resolution Timer. These standards are implemented in many of today's major browsers and are used by web developers to measure and improve the performance of their web apps. This talk presents the latest proposals for advancing the world of web performance.
Join us at the Evolution Bar in the Hyatt lobby after sessions on the last day of Velocity for a Closing Reception, sponsored by SOASTA. Make those final connections, and get the first peek at SOASTA's awesome new RUM service>
Ever notice the trajectory of emerging bands and technology startups are relatively the same? They both share a common goal - build a profession and lifestyle doing something immensely enjoyable that you wholeheartedly believe in effort to take the world by storm.
Dyn's Music Meets Tech events breaks the mold of traditional networking events by...
Akamai is serious about web performance and operations, but they still like to have fun. Join Akamai by the poolside to network with fellow Velocity attendees while enjoying great food, drink and entertainment.
Take a deep dive with us into the architecture and technology plus hear from one of our customers Kongregate about their experience with the technology in production delivering browser based games to over 15 million users.
In this session, we will take a deeper look into some of the great techniques delivered to us through the RESTful specification, applying them to developer API consumption practices with the intention of creating efficient best practices for rapid development. These are the lessons we learned as we reconstructed our API infrastructure at PayPal for our developer community.
MeetMe, Inc. is no stranger to growth. The leading global social platform is one of the most highly trafficked sites in the US, registering a new user every second — more than 86,000 a day. This growth represents a 300 percent increase over last year.
How do Google, Twitter, and Instagram ensure fast application performance at scale? One technique is asynchronous messaging using RabbitMQ to prevent application bottlenecks. In this session, we'll cover common asynchronous messaging patterns and how to implement them in RabbitMQ, common pitfalls to avoid, and how to cluster RabbitMQ for increased scalability and reliability.
Salesforce.com has developed mobile apps using HTML5, iOS & Android. It is critical to keep these apps performant across frequent deployments to production. Come learn how Salesforce built a mobile performance lab to keep our apps fast and prevent crashes on multiple platforms and devices. Learn the tools and techniques we use to setup your own mobile performance lab !
QuickBooks Online is the no. 1 small business cloud accounting solution worldwide. In this session we will talk about how we built a highly scalable, global SaaS offering and the lessons learnt along the way.
This presentation will look at the application delivery controller’s effect the development, deployment and maintenance of applications and web applications in particular.
The days of complete control are over. Today, the website is your organization's most important online asset and it is more exposed than ever before. Customers are measuring you, competitors are benchmarking you, and things outside of your control are robbing your site of performance.
AppDynamics Tech Evangelist, Steve Burton and Edmunds.com Senior Director of Production Engineering, John Martin, co-present this talk about DevOps culture and success.
In the old world of IT, if you didn't have hardware capacity or the budget to buy more, your project was dead in the water. Budget constraints can leave some of the best, most creative and most ingenious innovations on the cutting room floor. It’s a true dilemma for developers and innovators – why spend the time creating, when a project could be abandoned in a blink? That was the old world.
Continuous integration is critical to get quality mobile apps in the hands of customers on the device of their choosing quickly. This talk will reveal key considerations for evaluating third-party vs. custom-built solutions, sharing how Intuit created a homegrown virtual device lab and automated testing to enable continuous integration that reduced mobile app test time.
Learn how CoreSite is bringing together best-of-breed partners through the Open Cloud Exchange resulting in public, private, and hybrid cloud interconnection and management as well as connectivity to AWS direct connect.
The core infrastructure powering the web has seen significant performance improvements over the last few years. The browsers, devices, and networks are all much faster than they used to be, but has that resulted in a faster user experience?
In this talk, I will try to answer that question using real user measurement data from Google Analytics and Youtube.
A nearly infinite array of variables can affect performance. Gaining control over these variables allows providers and enterprises to introduce a greater degree of consistency and high performance into any environment. How can this be done? As infrastructure morphs into code, what was once secret becomes revealed.
Can your application scale with linear performance? Capacity planning for a Web or Mobile applications can be challenging but it is critical to any successful application launch. Testing how much load you can handle is just one part of it. Understanding where the bottlenecks are and what to do about it is just as important.
There is great progress on making web pages fast. Performance Management for native mobile applications is still in it's infancy. What if there was a free technology to help you solve the performance problems in your mobile applications?
Delivering truly great Web performance has never been easy. In today’s world, the challenge has grown exponentially. Akamai’s Chief Architect, Stephen Ludin, will highlight ways to identify the end users’ situation, determine the best optimizations to employ and how those optimizations directly impact performance – with the end goal of providing an answer to the situational performance challenge.
Building an efficient and scalable API design can mean the life or death of your developer community. While technical efficiency is a key factor in proper optimization, following a pragmatist approach to specification implementation and security can lead your API design down a path to certain doom.
In this session we'll take a deep dive into the technical design and optimization criteria to help us meet these goals. As a bonus, we'll demo some existing and upcoming tools we've been working on at Google to automate the process.
In this step by step session we set up Real User Monitoring (RUM) for a native mobile app from scratch. We will analyze the behavior of the application across a variety of aspects including performance metrics, user behavior data as well as crash reports. All tools used in the session are available for free.
At Velocity, we all share a common interest in making our Web and mobile sites robust and reliable for users. For some, this may be an altruistic pursuit. But for most, it's because we know our business grows when customers can quickly engage and transact. So how do we ensure that these transactions are performing their best for the users experiencing them?
For web performance professionals, this new age means having comprehensive management perspectives that include real user measurements. Join us as we examine the three reasons why RUM needs to be part of your management arsenal so you can make it easier to make the hard cash.
Is your web monitoring out of balance? Are you giving enough attention to API, REST and SOAP – especially as companies expand the web application partner channels and mobile applications? Join us on Thursday, June 20 at 2:00 p.m., as Dell product expert John Spirko leads a panel of web professionals who manage a very popular, very busy website.
In Japan, there are a lot of large scale web services that process >30 billions PV per day. For keeping such big services stable, Japanese engineers developed useful tools for web operation. Although these tools are popular in Japan, I think ops engineers in other countries do not know about that. In this talk, I will introduce such useful tools with use cases as many as possible.
FluxCapacitor is a github project that integrates all Netflix OSS Components with Cassandra, Graphite, Netty, and Jetty to provide a reference application for a highly-scalable, highly-available distributed system running on bare metal or AWS.
In this ignite talk, we will examine the popular science fiction series - Star Trek: The Original Series and show the audience what we learn from it to improve our engineering operations and practices.
Long-time IT infrastructure investor and startup mentor Sunil Dhaliwal explains how and why software developers and IT ops professionals are the driving force behind re-imagining IT infrastructure as we know it.
Everyone who has watched a Velocity presentation knows the importance of
monitoring our systems to ensure they're performing effectively. Many of the
best teams go further and link their metrics to business goals to ensure the
improvements they're making have a real impact on the bottom line. But despite
this work the question remains: are we measuring the things that really
Is it fast? Or do you just think it's fast? Does it matter?
There's a lot of talk about women in tech right now, which is a good thing. But instead of talking about what's wrong with tech, I want to talk about what's awesome about mountain biking, and how the tech community can learn from my personal experiences in that very different (yet strangely parallel) world.
We all use load balancers for various reasons – HA, Scaling, DDoS protection, Optimizations and many more. But to turn your network real estate into an Application Tap is something new and a rapidly growing trend that we will discuss in this session.