> Caching is based on file hashing. Not filenames.

I'm not really sure what this means, and I've used all the major CDNs/reverse proxies =)

Cloudflare for instance is generally controlled by the filetypes, while cloudfront is generally controlled by the cache control headers sent by the origin server.

Either way, I think exactly how it works needs to go into its own fully-fledged article. It's probably the most important thing a reverse-proxy does, and should be documented very clearly.

BTW one unique feature of cloudfront you might want to support, is setting cache control with vary on cookie. So with cloudfront, I can make non-logged in users see a cached page, while logged-in users see a non-cached page. It allows me to take a huge amount of load off the origin server, with no real downsides (in my case).