I've assumed that without SETTINGS, you couldn't really send a request anyways, so that was basically necessary HoLB. This introduces another frame that can cause additional HoLB. Though you could easily make the argument that we'd expect SETTINGS and the first MAX_PUSH_ID frame to be sent in the same packet. On Wed, Aug 9, 2017 at 5:11 PM, Martin Thomson <martin.thomson@gmail.com> wrote: > On 10 August 2017 at 08:00, Jana Iyengar <jri@google.com> wrote: > >> I think it's useful to avoid HoLB since the first request and the >> MAX_PUSH_ID frame may get reordered and/or the MAX_PUSH_ID frame may need >> retransmission. Push is already weak on benefits, so (without any evidence) >> I wonder whether any benefits due to it would be sensitive to this small >> variation. >> > > I was going to make exactly the same point, but then I realized that you > can't push without SETTINGS anyway and that has the exact same properties. > >
Received on Thursday, 10 August 2017 00:21:23 UTC