Re: HTTPS 2.0 without TLS extension?

Zhong Yu <> Tue, 23 July 2013 18:48 UTC

Return-Path: <>
Received: from localhost (localhost []) by (Postfix) with ESMTP id 2219111E830D for <>; Tue, 23 Jul 2013 11:48:44 -0700 (PDT)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: -9.149
X-Spam-Status: No, score=-9.149 tagged_above=-999 required=5 tests=[AWL=1.149, BAYES_00=-2.599, MIME_8BIT_HEADER=0.3, RCVD_IN_DNSWL_HI=-8]
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id IG-gNCJ7uCOF for <>; Tue, 23 Jul 2013 11:48:37 -0700 (PDT)
Received: from ( []) by (Postfix) with ESMTP id 556B711E836D for <>; Tue, 23 Jul 2013 11:48:36 -0700 (PDT)
Received: from lists by with local (Exim 4.72) (envelope-from <>) id 1V1hcI-0007SH-9t for; Tue, 23 Jul 2013 18:47:22 +0000
Resent-Date: Tue, 23 Jul 2013 18:47:22 +0000
Resent-Message-Id: <>
Received: from ([]) by with esmtp (Exim 4.72) (envelope-from <>) id 1V1hc9-0007N4-9e for; Tue, 23 Jul 2013 18:47:13 +0000
Received: from ([]) by with esmtps (TLS1.0:RSA_ARCFOUR_SHA1:16) (Exim 4.72) (envelope-from <>) id 1V1hc8-0002rW-AH for; Tue, 23 Jul 2013 18:47:13 +0000
Received: by with SMTP id n12so7354796oag.22 for <>; Tue, 23 Jul 2013 11:46:44 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed;; s=20120113; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type:content-transfer-encoding; bh=VXCsb3BRL/EskIqp9ifCi+G2+YKEOn/LhkLICVH2DHI=; b=XzCDa/D4oqTreRTLGVsyPfZBMTKmFyY/KObnyCI3Q5VrSEorZryXZ71TeIZ1DetsJx vA41eWr79KmtJvG7PuqOr5eHwyGDDPN8hFJtYa7Vmqg9Fsxl/OZcBctMuHTEnVfbFtdd 3Dd53oqfTBtKnJUVoiyCc+NBq3EIB49tnXT2hK2IrDSX/TUVSzC3OaKmt+F2fl5Rt5EH WSWvIiDuEHcxZt7SzYFWYypRaXaGG9S+RPh9NBq1Vu0NG4xNl10yh9ur2MzoKSLk8nBR xU6Tkh2u1bu4Jiq7g8hVO8mN7IcfpNaDYXqjkhib06fXf3JBdsGKKWY7IxlTqFo5KITI O1FA==
MIME-Version: 1.0
X-Received: by with SMTP id ml4mr32124907oeb.47.1374605204351; Tue, 23 Jul 2013 11:46:44 -0700 (PDT)
Received: by with HTTP; Tue, 23 Jul 2013 11:46:44 -0700 (PDT)
In-Reply-To: <>
References: <> <>
Date: Tue, 23 Jul 2013 13:46:44 -0500
Message-ID: <>
From: Zhong Yu <>
To: =?UTF-8?B?V2lsbGlhbSBDaGFuICjpmYjmmbrmmIwp?= <>
Cc: HTTP Working Group <>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Received-SPF: pass client-ip=;;
X-W3C-Hub-Spam-Status: No, score=-3.5
X-W3C-Hub-Spam-Report: AWL=-2.666, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, FREEMAIL_FROM=0.001, RCVD_IN_DNSWL_LOW=-0.7, SPF_PASS=-0.001
X-W3C-Scan-Sig: 1V1hc8-0002rW-AH aa9ed2bbc29f10fd90713866ebbf1c09
Subject: Re: HTTPS 2.0 without TLS extension?
Archived-At: <>
X-Mailing-List: <> archive/latest/18883
Precedence: list
List-Id: <>
List-Help: <>
List-Post: <>
List-Unsubscribe: <>

I agree TLS-ALPN is much better than TLS-Upgrade, but it'll take the
rest of the world some time to get there. Allowing TLS-Upgrade
meanwhile probably will not lessen the motivation to deploy ALPN,
since it's in the best interest of all clients and servers to reduce a
round trip.

*If* the spec allows TLS-Upgrade, some servers will use it before they
can do ALPN, and some clients will support it. Chrome will face the
heat - a competitor browser can talk to a server in HTTPS/2.0 with
Upgrade, yet Chrome can only talk to it in HTTPS/1.1.  Will Chrome
stick to its principle and refuse to speak with the vulgar Upgrade? I
bet a lunch that it will budge.

Therefore if the spec allows TLS-Upgrade, it might as well mandate it.

The other option is to absolutely forbid TLS-Upgrade and disown any
implementation that does it, deliberately or accidentally (the latter
being more likely).

Zhong Yu

On Tue, Jul 23, 2013 at 12:34 PM, William Chan (陈智昌)
<> wrote:
> FWIW, it seems reasonable to me to have the spec allow HTTPS 2.0 without TLS
> extension. If you want to Upgrade, be my guest. I have no plans for my
> browser to support that, and I don't think Google servers will support it
> either, because we care strongly about the advantages of TLS-ALPN vs
> Upgrade.
> IIRC, Twitter doesn't use NPN for the same reasons (lack of TLS extension
> support on certain mobile clients). I believe they don't care about public
> interop though, they just use dedicated VIPs with clients they control.
> On Mon, Jul 22, 2013 at 5:06 AM, Zhong Yu <> wrote:
>> The draft mandates TLS extension ALPN for any https 2.0 connections,
>> but why is that necessary? Why can't we also establish an https 2.0
>> connection through the Upgrade mechanism, without ALPN? TLS extension
>> may not be available/convenient on some platforms for some time;
>> requiring it may discourage some potential implementers.
>> Zhong Yu