Re: usability of 100-continue, was: HTTP2 Expression of Interest : Squid

On Fri, Jul 20, 2012 at 12:52 AM, Willy Tarreau <w@1wt.eu> wrote:

> [snip]
> > >
> > >    POST /public HTTP/1.1
> > >    Host: www.example.com
> > >    Expect: 100-Continue
> > >    Content-Length: 100
> > >
> > >    GET /private HTTP/1.1
> > >    Host: www.example.com
> >
> > In the above example the use of content-length requires that the server
> > drain 100 bytes regardless of what the client does.
>
> [snip]


So this would be yet another argument in favor of spdy-style framing and
multiplexing

Consider, for instance,

Client                  Server
  |                       |
  |======================>|
  | 1) SYN_STREAM         |
  |   id=1                |
  |   method=POST         |
  |   expect=100          |
  |   content-length:100  |
  |                       |
  |======================>|
  | 2) SYN_STREAM         |
  |   id=3                |
  |   method=GET          |
  |                       |

Due to the multiplexed framing model, the client can proceed with Stream #3
without waiting for the server to respond to the POST and without worrying
about whether the server will drop the data on the floor due to the
content-length header. If the server takes too long to respond to Stream
#1, it can attempt to go ahead and send the Stream #1 Data frames or it can
abandon the stream with a RST_STREAM without requiring the connection to be
killed off.

That said, fwiw, HTTP/1.1 clients are not supposed to be pipelining
non-idempotent methods anyway, so sending a GET on the same connection as a
POST, especially one that is using Expect: 100 is simply bad form.

- James


> Regards,
> Willy
>
>
>