“Edge computing, that’s the icing on top of the cake. That’s not what people wanted to start with. What they really wanted was just a simple platform where they could write code, not think about it, and it just worked,” Cloudflare CEO Matthew Prince told The New Stack. “What we got wrong three years ago, and I think what almost every edge computing platform has gotten wrong over the years, was they lead with speed. That’s, again, the icing on top, it’s not what a really great serverless platform is.”
Instead, Prince said, what teams and developers really want from serverless is cost efficiency, ease of use, and most importantly, compliance.
“The thing that is actually even more important than ease of use is actually compliance. Not very sexy, it’s not very fun to talk about, but if your general counsel or your CIO says you must do X or you can’t do Y, that really ties the developers hands,” said Prince. “So compliance is a really important piece. And I actually think that that’s going to be the killer feature of any edge computing platform going forward.”
Whereas Cloudflare Workers were limited with certain restrictions regarding things like CPU time, memory usage, and program size, Workers Unbound are left comparatively unrestrained, with unthrottled CPU usage, automatic scaling, and the ability to “run heavy workloads without having to worry about overly restrictive CPU constraints.” This allows users to perform more complex tasks than with Workers, while at the same time spending less. According to the company, Workers Unbound, for the same workload, “can be 75 percent less expensive than AWS Lambda, 24 percent less expensive than Microsoft Azure Functions, and 52 percent less expensive than Google Cloud Functions.”
Prince explained that there were a couple factors that made Workers Unbound so much cheaper than its competitors.
“We got to start from scratch and we really built this to be just in a hyper-efficient computing platform. Turns out, if you can learn from everyone else’s mistakes, you get an advantage there,” said Prince. “The second thing, which is the most important, is that we really built a more efficient way of running code. Where traditional serverless platforms are spinning up VMs or some sort of a container that has an enormous amount of overhead, we took a different approach. We used a technology called an isolate. An isolate is a different form of sandboxing than a VM or a container that is much, much more lightweight.”
Prince said they built this using technology from the Chrome project, using a “heavily modified version of V8” that can input from or output to any TCP/IP application.
As for the “killer feature” of compliance, Prince argues that Cloudflare’s ability to locate all of these workloads across its vast global network puts it in the position to achieve compliance with local regulations regarding user data and processing.
“Because Cloudflare’s network today spans over 200 cities in more than 100 countries worldwide, we are already in the places where those governments are writing the rules that say local data has to stay local. The workers platform allows you to ensure that you can write code in one place and then we will help ensure that you comply with that increasingly complicated regulatory environment. We think that that’s going to be the killer feature for edge computing going forward. It’s what our largest customers are asking us for. We have the beginnings of tools to make sure that we are complying with that, but my hunch is that over the next three years, that’s going to become more and more of a critical feature for any computing platform,” said Prince.
At this time, The New Stack does not allow comments directly on this website. We invite all readers who wish to discuss a story to visit us on Twitter or Facebook. We also welcome your news tips and feedback via email: firstname.lastname@example.org.