Return-Path: X-Original-To: quic-issues@ietfa.amsl.com Delivered-To: quic-issues@ietfa.amsl.com Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 69E89120143 for ; Mon, 25 Mar 2019 15:51:22 -0700 (PDT) X-Virus-Scanned: amavisd-new at amsl.com X-Spam-Flag: NO X-Spam-Score: -8.001 X-Spam-Level: X-Spam-Status: No, score=-8.001 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIMWL_WL_HIGH=-0.001, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, HTML_MESSAGE=0.001, MAILING_LIST_MULTI=-1, RCVD_IN_DNSWL_HI=-5, SPF_PASS=-0.001] autolearn=ham autolearn_force=no Authentication-Results: ietfa.amsl.com (amavisd-new); dkim=pass (1024-bit key) header.d=github.com Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id xiZ-DREOWpvZ for ; Mon, 25 Mar 2019 15:51:20 -0700 (PDT) Received: from out-7.smtp.github.com (out-7.smtp.github.com [192.30.252.198]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id A0EE6120134 for ; Mon, 25 Mar 2019 15:51:20 -0700 (PDT) Date: Mon, 25 Mar 2019 15:51:19 -0700 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=github.com; s=pf2014; t=1553554279; bh=sbaP7xSSmbWBBw3rchF16iDFfJHI3q+gFAo37UhorEk=; h=Date:From:Reply-To:To:Cc:In-Reply-To:References:Subject:List-ID: List-Archive:List-Post:List-Unsubscribe:From; b=ES2jv2Q9C4u5cS0x/vAHyIOXxb23qEif0sO47vP+MwkazoBA+9TATgX4VwC2f8o5t khCixMLXmajVCRmM2RXvaAnl8MI7RlK6fd8kFDrKAIFHNfst5FH4ZnQZbVJIQXHhVS TS+eRBUtLQySn4A3ROVbhL3y56GSNmEM9iAG002c= From: ianswett Reply-To: quicwg/base-drafts To: quicwg/base-drafts Cc: Subscribed Message-ID: In-Reply-To: References: Subject: Re: [quicwg/base-drafts] kInitialRtt of 100 msec is too aggressive (#2184) Mime-Version: 1.0 Content-Type: multipart/alternative; boundary="--==_mimepart_5c995b679999d_39c73f8d8f4d45c0444415"; charset=UTF-8 Content-Transfer-Encoding: 7bit Precedence: list X-GitHub-Sender: ianswett X-GitHub-Recipient: quic-issues X-GitHub-Reason: subscribed X-Auto-Response-Suppress: All X-GitHub-Recipient-Address: quic-issues@ietf.org Archived-At: X-BeenThere: quic-issues@ietf.org X-Mailman-Version: 2.1.29 List-Id: Notification list for GitHub issues related to the QUIC WG List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 25 Mar 2019 22:51:23 -0000 ----==_mimepart_5c995b679999d_39c73f8d8f4d45c0444415 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Thanks for the additional information. The data is from 2009(https://www.ietf.org/proceedings/75/slides/tcpm-1.pdf), but still seems largely plausible in terms of RTTs, though I'm sure some details have changed. One thought is that the client may have a better idea of what it's RTT is than the server does when there is no resumption information available. As such, we could pick a fairly conservative value(ie: 1s) on the server and if the client retransmits(and the server receives) the ClientHello more frequently, the server should retransmit its response as well. In practice, the current anti-amplification limit of 3x amount to exactly that in a lot of circumstances, so making it core to the handshake mechanism may be wise. On the other hand, this is a special case unless we tweak the anti-amplification limit to be min(3 packets, initial server flight). The 4 second timeout I mentioned is for the handshake. I would expect live migrations to not occur during the handshake, or at least not expect the handshake to complete in that case, but maybe that's not true? We experimented with bumping the timeout to 10 seconds a while back in gQUIC, but there were some performance regressions and we didn't dig in deeply enough to understand why. -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/quicwg/base-drafts/issues/2184#issuecomment-476407605 ----==_mimepart_5c995b679999d_39c73f8d8f4d45c0444415 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable

Thanks for the additional information. The data is from 2009(= https://www.ietf.org/proceedings/75/slides/tcpm-1.pdf), but still see= ms largely plausible in terms of RTTs, though I'm sure some details have = changed.

One thought is that the client may have a better idea of what it's RTT= is than the server does when there is no resumption information availabl= e. As such, we could pick a fairly conservative value(ie: 1s) on the ser= ver and if the client retransmits(and the server receives) the ClientHell= o more frequently, the server should retransmit its response as well. I= n practice, the current anti-amplification limit of 3x amount to exactly = that in a lot of circumstances, so making it core to the handshake mechan= ism may be wise. On the other hand, this is a special case unless we twe= ak the anti-amplification limit to be min(3 packets, initial server fligh= t).

The 4 second timeout I mentioned is for the handshake. I would expect= live migrations to not occur during the handshake, or at least not expec= t the handshake to complete in that case, but maybe that's not true? We = experimented with bumping the timeout to 10 seconds a while back in gQUIC= , but there were some performance regressions and we didn't dig in deeply= enough to understand why.

&m= dash;
You are receiving this because you are subscribed to this thre= ad.
Reply to this email directly, view it on GitHub,= or mute the thread.3D""

= ----==_mimepart_5c995b679999d_39c73f8d8f4d45c0444415--