Re: CONTINUATION was: #540: "jumbo" frames

In message <CABkgnnWpkmT0SxENxEEAhjVRSGQ+G_cBrCBM1GKoSZ6Wo6CFDg@mail.gmail.com>
, Martin Thomson writes:
>On 26 June 2014 15:41,  <K.Morgan@iaea.org> wrote:
>> Please share your use case for CONTINUATION when the header block is not >16KB.
>
>This would be a great line of argument if we were creating a new protocol.
>
>Sadly we are not.  We are responsible here for accommodating the uses
>of HTTP that exist.  We're not dictators, we're curators.

"the uses" or only "some of the uses" ?

Is high performance load-balancing in the set of uses "we are
responsible for accomodating" ?

>As such, we can't unilaterally impose restrictions like this.

My proposal for jumboframes contain no restrictions that differ from
restrictions in HTTP/1.

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.

Received on Friday, 27 June 2014 06:05:13 UTC