One of the biggest question that comes in is how is old browsers going to deal with HTTP/2 web servers and vice-versa.
HTTP/2 uses an Upgrade and Discovery method to find if a server supports HTTP/2 or not.
An HTTP/2 client never makes a HTTP/2 request by default. It always makes an HTTP/1.1 request with
Upgrade: HTTP/2.0 header. If the web server is HTTP/2 capable then the server responds with
HTTP/1.1 101 Switching Protocols status. But if the server is not capable of HTTP/2 then it simply returns an HTTP/1.1 response.
Clients usually remember whether the server supported HTTP/2 or not. And then the next time it makes an HTTP request it tries to make an HTTP/2 request directly instead of relying on upgrade round trip again.