Re: HTTP/2.0 Magic

On 1/02/2013 2:24 p.m., Martin Thomson wrote:
> The conclusion that we reached in the interim was that no matter how
> HTTP/2.0 was started, there would be some magic that started the
> session.

I have been working on this for a while now in the background. Lets just 
say that utilizing the first 16 bits of any connection it is possible to 
uniquely identify HTTP/1, SPDY v2 and SPDY v3 and WebSockets clients. I 
am still investigating the TLS, SSH and SNI bit patterns to see whether 
we can fold port 443 and port 80 together for HTTP/2-enabled servers.

Magic will not be a big issue and we can avoid worrying about generating 
some BOM magic octets just yet. The SPDY style of versioning still being 
used in the HTTP/2 Draft document provides sufficient binary variation 
to detect sub-versions as things progress and to become the magic octet 
if future framing versions require one.

Amos

Received on Friday, 1 February 2013 10:07:54 UTC