Webmaster Track

Capacity Planning for LAMP Installations

John Allspaw, Flickr

This talk will focus on the "do or die" fundamentals of capacity planning for a site that will experience rapid growth. We'll cover some of the techniques and tools used to measure and predict capacity needs, and considerations for carving up your back-end architecture so that capacity and operations management can scale alongside your application.

Measuring Success: Using Web Analytics to measure ROI

Alan K'necht, K'nechtology, Inc.

This will be a product neutral presentation; however since K'necht's primary tool for delivering Web analytics to his clients is WebTrends, screen shots presented will be obtained from using WebTrends. Attendees will learn that there is more to Web analytics than simply counting hits and visitors. K'necht will dive into the new breed of Web analytic tools that allow users to measure the success of changes to a Web site, search engine optimization efforts and online and off-line marketing efforts.

Specific areas to be covered include:

• Saying goodbye to hits
• What measurements should you be using?
• How to effectively track unique visitors
• The importance of separating paid from organic search engine traffic
• The value of persistent cookies
• Pros & cons of traditional log files vs. hosted data collection solutions

Real life examples of how this theory has been applied and adapted will be presented as well.

Networking for Large Applications

Nick Heyman, Facebook

Large Web applications present new challenges for network design and deployment. Content load balancing is playing a bigger and bigger role in the design of large architectures, but little is understood about the best way to deploy redundant routing and balancing. This session will take a look at some of the issues involved with huge volume Web site networking and discuss potential implementations and deployments.

Real World Web Performance and Scalability

Ask Bjoern Hansen, Develooper LLC

Learn how to build your overall architecture to beat the pants off yesterday's micro-optimizations:

• Horizontal scaling: the geek girl next door
• Caching here, there and everywhere
• MySQL configuration
• Table types and tweaks
• Replication strategies
• When to use what where
• How to avoid scaling vertically on the backend
• Partition your data the right way when your database can't keep up
• The when, where, and what about session data
• Offload light jobs to light processes
• Manage your resources to get the most out of your hardware
• Make a job queue to gracefully deal with big traffic peaks
• Tips for dealing with logging

Beyond the Filesystem: Designing Large Scale File Storage and Serving

Cal Henderson, Flickr

Flickr stores over a billion photo files, serving around five billion image impressions every week. In this session, Flickr Architect Cal Henderson will talk about the challenges faced designing and building affordable storage systems for large scale file serving operations. The session will focus on the various problems, possible solutions, the available hardware and software technologies, and some real-life case studies.

Configuration Management for Clusters with Subversion

Kevin Murphy, ex-Friendster

If you've ever been in the middle of a crisis and found yourself asking, "What changed? Everything was working fine on Tuesday!" you understand the need for configuration management. One of the primary outputs of Unix administrators is carefully-tuned configuration files. Developers use version control to keep track of changes for the entire lifespan of a project — you can use the same tools to track the state of your running systems. This presentation will talk about the whys and hows of using a version control system to manage configuration profiles for clusters. An implementation example using Subversion and Python will be provided.

Navigating the Web Hosting Maze

Erica Douglass, Simpli Hosting

There are thousands of hosting companies out there, offering plans ranging from a few dollars a month to several thousand dollars a month. How do you pick the right hosting company for your application? How do you ensure that you won't get ripped off by a fly-by-night operation or spend thousands of dollars more than you need to? Erica Douglass, 8-year web hosting industry veteran and owner of a successful hosting company, guides you through picking the right hosting company and plan for your application in a refreshingly honest, straightforward manner. Questions welcome!

Security 2.0

Chris Shiflett, OmniTI

Web 2.0 has been described as many things. It's the Web as a platform, a network of networks, the architecture of participation. However you choose to define it, the way we build applications online has changed. Web sites do more by empowering users, but this has opened a Pandora's box. Cross-site scripting (XSS), cross-site request forgeries (CSRF), and Ajax are being combined in creative new ways to launch sophisticated attacks that penetrate firewalls, target users and spread like worms. This talk examines this new threat, dubbed Security 2.0, by demonstrating some hypothetical and real exploits as well as discussing methods of safeguard and prevention.

Challenges of Large Scale Hosting

Chris Lea, Media Temple

Hosting Web sites is a challenging proposition in the current Internet landscape. Rapidly evolving technologies coupled with increasing customer needs and expectations force hosting companies to continually make challenging decisions about their service offerings. These challenges are amplified as the number of hosted sites increases, meaning that larger hosting companies face unique obstacles that smaller hosters are often not troubled by.

In my talk, I will discuss how larger hosting companies think about hosting. In particular there will be a focus on the technical issues and concerns that are problematic in an environment supporting tens of thousands of customer accounts.

Topics discussed will include:

• Support: Staffing, training and availability
• Email: SPAM, blacklisting, support and SLAs
• Storage: Disk IO strain, customer demand, backups
• Network: External and internal attacks, bandwidth vs. transfer, security
• Software: Choosing what to support, "bleeding edge" vs. stability, security
• Virtualization: The increasing role of virtualization in hosting, benefits to customers

Monitoring Large Web sites — Why is Your Site Slow?

Artur Bergman, Six Apart / Livejournal

Knowing what your site is doing is a hard problem, and it becomes more difficult, yet more important, the larger your site grows. How do you monitor and instrument large Web sites? How do you track requests through the entire system, from front-end load balancers to the database? Where in this path is your bottleneck? This session will look into tools such as ganglia, nagios, cacti and custom scripts that let you get graphical representations of your systems. Learn how you can get more relevant data out of your app and feed them into the visualization tools to better understand what your system is doing and where it is going wrong.