Re: BLOCKED frame specification

• From: David Krauss <potswa@gmail.com>
• Date: Wed, 9 Apr 2014 16:53:29 +0800
• To: Yoav Nir <ynir.ietf@gmail.com>
• Cc: Martin Thomson <martin.thomson@gmail.com>, Mark Nottingham <mnot@mnot.net>, Hasan Khalil <mian.hasan.khalil@gmail.com>, HTTP Working Group <ietf-http-wg@w3.org>
• Message-Id: <02462911-B0FE-47F3-B9AD-A6D651CFB4D7@gmail.com>

On 2014–04–09, at 4:47 PM, Yoav Nir <ynir.ietf@gmail.com> wrote:

> I want the information that it conveys, and this seems like a workable way of providing it.  I don’t think a RST_STREAM code or GOAWAY data can replace this, because they will only occur after a timeout. The BLOCKED frame can be sent after a much shorter time (1 RTT?) or even immediately.
> 
> On second thought, “debugging” is probably more important, because correct implementations should never reach this state.

Then why doesn’t PING suffice? You can send it at any time and there’s space for an implementation-defined payload value.

Received on Wednesday, 9 April 2014 08:54:11 UTC