Extend the power of your core cloud to the edge, and empower your teams to innovate. By moving data and applications as close to your end users as possible, you can deliver fast, highly personalized experiences to customers around the world. And now you can join the private beta for Compute@Edge, our new serverless compute environment built to take you further at the edge.
for 
Reddit needed to migrate their front end to an entirely new stack — all while beta testing a new user profile page on that stack. Because Fastly supports URL redirects from the edge with our powerful edge computing language, which speeds up response times and reduces origin server load, we were able to help them reach their goals, faster.
Using our edge computing language, they wrote code that executed at the edge to programmatically redirect the profiles of 75,000 users. Users were automatically redirected to the appropriate stack depending on their opt-in status, and Reddit was able to migrate the front end cleanly while providing users a seamless experience — without impacting performance.
for global Saas healthy living app and website
Companies with decades of history know the importance of digital transformation. One such company, which offers popular healthy living services online, turned to Fastly to help them migrate off their legacy CDN, and find a more efficient way to manage URL redirects.
With Fastly’s Edge Dictionaries, they could programmatically add new redirects whenever their marketing team wanted to send users to new parts of the site. On their legacy CDN, all redirect changes required a support ticket and a waiting period for professional services. Fastly’s solution was instantly configurable, too: the team could customize redirects, and create rules to quickly push them out or pull them back. With our API they were also able to integrate with their configuration management tool, creating a more agile process for launching new online experiences as they head into the future.
for 
HotelTonight’s success relies on the ability to share the best available hotels closest to their users. But because they rely on cellular GPS data, they often deal with incredibly specific latitudinal and longitudinal information that makes it difficult to group requests together from users who may be right next to each other. And they were inundated with the sheer number of requests to their origin, as different devices and apps generate different URLs.
But with our edge computing language, HotelTonight leveraged Fastly’s edge to create geo-fenced URLs, which grouped requests from users within the same proximity together. This gave them the accuracy they needed to serve users the best hotel rooms, while reducing the amount of unique requests to the hotel inventory API by several orders of magnitude. They were also able to use our edge computing language to rewrite URLs at the edge, so that non-standard URLs wouldn’t negatively impact cache hit ratio. With our edge compute technology, they were able to return results in under 50ms, with an 80% reduction in traffic to their origin.
for leading American airline
A leading international airline wanted to create a new online travel experience geared toward millennials. To do so, they needed to A/B test on a subset of production traffic to see what would perform well before rolling it out.
Using our edge compute technology, they set up the new site on several canary servers and created logic at the edge with Edge Dictionaries to control where backend visitors were sent. Since the test environment was contained, they didn’t have to worry about bad code negatively impacting their entire site. This allowed their developers greater agility, the ability to push out new tests quicker, and adjust on the go. Ultimately, they were able to make the data-informed decision to roll out their new site experience to their audience. And because their site was served from the network edge, they were able to deliver a better visitor experience, all while reducing traffic to origin.
for global travel and hospitality brand
One leading travel and hospitality business leveraged Fastly’s edge compute technology to accelerate A/B testing at the edge, and quickly uncover what was resonating most with their users. Through our powerful edge computing language, they bucketed customers into distinct groups with randomization, tested out new page layouts, and leveraged cookies to ensure users consistently received the same content. They did so while caching two variants of the same URL at the edge, thereby preserving their SEO rankings — whereas other solutions would cause them to maintain two separate URLs.
Once they identified the highest performing content, they rolled it out to a larger percentage of users, which boosted overall engagement. Serving this content from the edge reduced the burden on their servers, resulted in a better user experience, and let their team continue to iterate quickly and intelligently.
for internationally acclaimed magazine
A renowned, decades-running magazine wanted to provide more value to their paid subscribers, and entice non-paying customers to join a higher tier through offering membership-only content. Using Fastly’s Edge Dictionaries and client insights, they were able to build a paywall that runs entirely at the edge — without dealing with costly, time-consuming origin lookups to check a reader’s membership at every page visit.
When a reader requests access to gated content, Fastly checks the magazine’s authentication server for that reader’s status, caches it at the edge, and serves premium content from the cache. This ensures premium subscribers get the VIP experiences they expect — without any slow downs due to constant authentication lookups — and that paid content is only accessed by those authorized subscribers, which effectively protects intellectual property. And if a member cancels their subscription, Fastly instantly updates their status at the edge, blocking subsequent requests and ensuring everything runs smoothly.
for international daily newspaper
Feature flags are a well-known software deployment technique used to turn certain functionality on or off (without having to deploy new code), which allows teams to test new features while minimizing risk. An international daily newspaper wanted to test a new feature with a small subset of customers, and decided to implement feature flags leveraging Fastly’s powerful edge computing language.
This empowered developers to segment new deployments to a subset of customers, and test new versions in production before releasing them externally, both of which greatly reduced risk. As results started to come in, they decided to gradually dial up the percentage of users exposed to the new feature, until they reached 100% of customers. At that point, they were able to remove the feature flag and make it a permanent feature.
We’re pushing edge innovation even further with Fastly Labs. Your developers can explore our latest projects, and interact with experimental, in-progress ideas — all giving your teams an edge, and empowering them to step into the future, faster.
