We continue the conversation with Zack Bloom of Cloudflare. In this episode Zack and I talk about computing at the edge. He walks me through developments that accelerated at Cloudflare due to the needs of services like Optimizely.
Zack and I talk about how Cloudflare took a different approach when addressing the market for edge computing services. Cloudflare had customers who needed to quickly modify websites, edge computing made sense in this context.
API based systems at scale needed excellent performance. Edge computing supported these goals.
With Cloudflare's edge computing, the problem of cold starts turned into an opportunity to solve. Cloudflare took the underlying engine that runs Google Chrome and leveraged that for their edge computing solutions.
With Cloudflare Workers, your code can instantly scale effortlessly. The cost of the infrastructure is greatly reduced as well.
Through Workers there is an API into Cloudflare's caching system. This gives developers more flexibility with how they leverage Cloudflare's tools.
We get into the details of how Worker's get involved in the web request flow. We talk about how Cloudflare Workers occupy a privileged position in the flow of a request. They can intercept and alter a request in a trusted and secure manner that is also extremely fast.
We dive into questions of parsing HTML and Cloudflare's history of handling it. Workers have an HTML Rewriter using simple CSS selectors. Workers are also capable of different types of call-backs to databases or servers to get dynamic content.
We work through more examples of how people decide what to migrate into Workers.