- From: Jason Greene <jason.greene@redhat.com>
- Date: Mon, 2 Jun 2014 10:15:51 -0500
- To: Poul-Henning Kamp <phk@phk.freebsd.dk>
- Cc: "Jason T. Greene" <jgreene@redhat.com>, Roberto Peon <grmocg@gmail.com>, Greg Wilkins <gregw@intalio.com>, HTTP Working Group <ietf-http-wg@w3.org>
On Jun 1, 2014, at 3:13 PM, Jason Greene <jason.greene@redhat.com> wrote: >> 8-cores, 40gbit/s -> 4x10gbit/s anyone ? >> >> What *is* the highest rate anybody has processed HTTP/2.0 according >> to the current draft anyway ? >> >> I don't recall seing anybody brag about that yet ? > > It might be possible to keep the stateful delta operations and achieve http 1.1 throughput levels if the receiver, and not just the sender could negotiate the dynamic table size to 0. To clarify what I meant, as my wording is poor: The sender trumps the receiver until the sender processes the SETTINGS frame, allowing for a significant number of requests to be sent with a table size of 4096 (since headers have no flow control, and also the wording seems to not disallow a client from putting off processing of the inbound frame). This makes it hard to have a (mostly) stateless proxy that splits or aggregates requests, limiting the scalability and throughput of the system. -- Jason T. Greene WildFly Lead / JBoss EAP Platform Architect JBoss, a division of Red Hat
Received on Monday, 2 June 2014 15:16:30 UTC