Compute@Edge: insights from companies running traffic on our serverless compute environment
Unlocking performance, agility, security, and creativity without having to worry about underlying infrastructure has been the mission of our serverless compute environment — Compute@Edge — from its inception.
Since moving out of beta and into real production environments, we've been astounded by our customers’ brilliance as they continue to experiment and test, now with production traffic. Their use cases make the true promise of serverless tangible. And this is the case no matter the industry, from publishing to ecommerce, media to travel.
During our last Altitude customer conference, we heard first-hand experiences from three organizations. Vox Media, Mux, and PerimeterX have all been building on Compute@Edge — and are energized by its potential. Here are some of their takeaways.
Media manipulation at the edge for optimal performance
Media isn’t what it was a few years ago. Today, there are hundreds of types of smartphones, smart TVs, and set-top boxes, and content needs to work optimally on each format. This means customizing the video manifest for each device, as well as taking into account things like screen size and bandwidth. Unfortunately, going back to the origin for each of the many combinations of device and network negatively impacts performance.
Engineers from Mux, a San Francisco-based video technology start-up that powers video streaming for the likes of Fox and Vimeo, said being able to manipulate media at the edge is one of the game-changing benefits of our serverless compute environment. With Compute@Edge, the original manifest files are cached close to the user, and then customized versions for different devices are generated on-the-fly, resulting in fewer origin round-trips for optimal performance.
These performance levels are also attractive to Vox Media. The company customizes and translates loads of content for different outputs, personalizing both content and functionality. “Having logic and caching executed at the first point of contact is fantastic [from] a performance perspective,” Vox Media CTO Pablo Mercado told us, adding that the potential Compute@Edge has for unlocking creativity also excites his team.
A middleware stack that enables flexibility and agility
The team at Vox Media likes to think of Fastly not just as a content delivery network, but as a broader “business delivery network.” Every day, millions of people interact with podcasts, videos, and digital media content from the Vox portfolio of brands, including The Verge, New York Magazine, and vox.com — and all this traffic goes through Fastly.
Vox Media is currently prototyping on Compute@Edge, and their engineering team already sees it as a beneficial “middleware stack,” unlocking flexibility and agility across the entire business. Where inserting new logic upstream can sometimes be brittle, in this environment, developers, engineers, and creative thinkers can assert business logic or test out new functionality with increased agility. Whatever they build can be integrated with ease and safety — and with greater developer economics.
New capabilities for seamless security
The team at PerimeterX, which provides behavior-based bot protection, is excited that with Compute@Edge, they can enhance their security ecosystem by adding complexity and logic, without overcomplicating things. There is no need to add another application, and customers never need to change anything on their application backend to enjoy new capabilities. Running on the edge also means malicious traffic and activity is handled safely away from data centers and applications.
“We're leveraging the snappiness of configuration updates done by Fastly and the ability of code to write and read from edge dictionaries,” Ido Safruti, CTO of PerimeterX, told us. “That helps us push decisions and make it even more independent and on the edge, not requiring or eliminating some of these API code that we tended to use before.”
Bringing it all together
Hearing these experiences and insights is valuable for us because our roadmap for Compute@Edge is built hand-in-hand with our customers.
As we laid out in our product vision for 2021, we’ll continue our work toward general availability, and we’re also adding more languages and tools developers are familiar with in order to enable broad adoption of the platform by builders of all skillsets.
As more organizations discover the potential of Compute@Edge, we look forward to sharing even more stories and learnings, so we can build the future together. To stay abreast of Compute@Edge updates, sign up for our email list.
This blog post was based on a panel hosted at Altitude, our customer summit held in November 2020. Watch other talks on the future of Fastly and the internet from the virtual event here.