The Fastly team is attending Fluent this week in San Francisco. We’ll be at booth #206 right next to our friends from New Relic. Come by to talk with one of our engineers, grab a Fastly shirt, and learn how to win our DevOps survival kit.

Today, we’re excited to welcome three new Fastly team members.

In this, our final API Caching installment, we're going to explore how to use Surrogate Keys to reduce the overall complexity of caching an API.

At Fastly, we’re always working to make our systems faster and more reliable. One of the more difficult problems we’ve faced is efficient cache invalidation across our global network, or as we call it: Instant Purging. When content changes, our customers issue a purge request, which we then need to deliver to each of our cache servers. The system that handles these purge requests is codenamed Powderhorn.

Build a network that can scale indefinitely, be managed by a small crew of skilled ops and network engineers, and handle current web traffic and the next generation of protocols. Sounds impossible, right?

Not true.

When planning a major content delivery network, you’d think that it would make sense to put your equipment where the most people are, right?

Not always.

In Part 1, we covered the basics of using Fastly to accelerate a comments API. Using Instant Purge, we hooked into model callbacks to ensure that the appropriate content was purged from the cache whenever data changed. In this article, we’ll build upon the original approach and use one of Fastly’s more advanced features: cache control.

Over the last few months, we’ve completely overhauled Fastly’s entire logging infrastructure to keep up with our ever-increasing traffic. In addition to improving efficiency and reliability, we added features to make it easier for our customers to integrate with the logging providers they already use.

Last week, Fastly engineer Jason Cook spoke about TCP tuning at the 2014 Linux Conference in Australia. His session covers tuning several aspects of your application and the underlying TCP stack to deliver the best possible performance over the public Internet. If you missed it, check out the video below.

Since day one, we’ve baked the concept of versions right into our system at a fundamental level. We want customers to feel empowered to try new things and to be able to quickly roll them back or create a new version if changes are needed. We’ve brought DevOps best practices to our CDN.

During the holiday season, which normally accounts for 20-40% of total sales annually, it's critical for e-retailers to provide their users with a smooth, responsive shopping experience. Recent research has shown that 40% of online shoppers will move to a competitor’s site if yours doesn't load in 3 seconds or less. In fact, about $3 billion in revenue is lost annually due to customers abandoning shopping carts on slow web sites. With that in mind, here are Fastly’s top 5 tips for optimizing and speeding up your site for the holidays.

The web has come a long way since the 90s. In the past, sites were commonly driven by a single, monolithic application that acted as the only communication medium to a centralized database. The modern approach is to break this one large application into a set of interdependent and cooperative services.

Ensuring a smooth end user experience with short page load times - whether for shopping online, reading the news, or collaborating on software - is one of the main reasons Fastly customers use our Content Delivery Network. Viewers' expectations for page load times have become increasingly demanding.

Fastly Caching Software is built on Varnish Cache, an open source HTTP accelerator. We are big believers in open source - we provide free services for Ruby Gems, PyPI, Perl and Debian among others, and we’re regular contributors to Varnish, Chef, Perl and Ganglia. Fastly fully supports VCL (Varnish Configuration Language), allowing customers to both download Fastly-generated VCL and upload custom VCL to mix and match with the Fastly configuration. VCL offers our customers unmatched control over request routing and on-the fly request and response modification.

At Fastly, we’re constantly working to upgrade our network and expand global capacity. Last week, we built up capacity in our Ashburn, New York City, and Los Angeles POPs and brought a new Miami POP online to improve user experience in South and Central America. We also brought a second London POP online and increased the capacity in our existing London and Frankfurt POPs. In addition to our existing POPs in Tokyo and New Zealand, we’ve added new POPs in Singapore and Hong Kong to bring our customers’ content even closer to users in Asia Pacific. These upgrades are already available to current customers.

At Fastly, we’ve been steadily building the world’s most advanced CDN. Since our founding, we’ve expanded the public network to include 17 globally distributed points of presence (POPs) and are now serving more than 20 billion requests per day.

For the majority of the population, Thanksgiving and the rest of the holiday period are a time for family, feasting and frivolity. For eCommerce managers, however, it's the busiest period of the year with almost a quarter of yearly online sales happening over a 4 week period.

We're proud to have been a part of New Relic's first FutureStack conference (and equally proud to have New Relic as a Fastly customer)! If you didn't make it to FutureStack, you can watch Fastly CEO Artur Bergman's talk below. We hope to see you there next year.

We’re headed to London this week for Velocity Europe! If you’re attending, please stop by kiosk #306 in the Exhibit Hall to meet the Fastly team and chat with our engineers.