Re: Rechartering HTTPbis

On 24/01/2012 9:48 p.m., Willy Tarreau wrote:
> Hi Poul-Henning,
>
> On Tue, Jan 24, 2012 at 08:13:20AM +0000, Poul-Henning Kamp wrote:
>
>>> This work will be known as "HTTP/2.0", unless the Working Group
>>> determines that this isn't suitable (e.g., for interoperability).
>> I really don't think it qualifies for that pretentious name, because
>> all it is set up to be, is tunneling HTTP/1.1 more efficiently
>> through the tubes.
>>
>> In my mind, the effort sketched out would be correctly titled
>> "Beatify the SPDY protocol as a carrier of HTTP/1.1 traffic"
>
> HTTP/1.1 has a number of issues that make the current spec very
> heavy and implementations complex (eg: remember you can't fold
> set-cookie, the issues with multiple content-length, etc...).
> Taking the opportunity of a new version to clear a few of these
> old issues would be nice.

Not to mention the issues with Host:, CONNECT-vs-Upgrade:,


>
>> HTTP/2.0 would in my mind be an attempt to actually improve the
>> protocol, possibly going as far as replacing everything but the
>> first line which we would have to keep, to ensure the ability
>> to interoperate with earlier HTTP protocols)
> Maybe that's an option too.

One I hope we pick up. Remember, *semantics* is the only thing required 
to be preserved when doing so. A short and streamlined for of transmit 
version of the details above can be found I'm sure, and "messing them up 
again" for transmission over HTTP/1.1 hops is an option.

For example, simply truncating the common header names down to 1 or 2 
bytes and moving to a better timestamp format we could meet all the 
HTTP/2.0 requirements:
  * chop out a visible % of HTTP traffic size
  * be syntactically incompatible with HTTP/1.x
  * capable of trivially gatewaying 1.1-over-2.0 and vice versa
  * with full semantic and feature equivalence
  * and lower barriers to implementation (fingers on telnet + fewer 
bytes = win)

AYJ

Received on Tuesday, 24 January 2012 10:09:48 UTC