Re: [tcpinc] WGLC for draft-ietf-tcpinc-tcpcrypt - draft header

Daniel B Giffin <> Thu, 16 February 2017 02:19 UTC

Return-Path: <>
Received: from localhost (localhost []) by (Postfix) with ESMTP id E8BCF129C55 for <>; Wed, 15 Feb 2017 18:19:43 -0800 (PST)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: -1.902
X-Spam-Status: No, score=-1.902 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, RP_MATCHES_RCVD=-0.001, SPF_PASS=-0.001] autolearn=ham autolearn_force=no
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id puaaHhuJxPsf for <>; Wed, 15 Feb 2017 18:19:42 -0800 (PST)
Received: from ( [IPv6:2001:470:806d:1::9]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by (Postfix) with ESMTPS id 40613129C51 for <>; Wed, 15 Feb 2017 18:19:42 -0800 (PST)
Received: from (localhost []) by (8.15.2/8.15.2) with ESMTP id v1G2JeQI019138; Wed, 15 Feb 2017 18:19:40 -0800 (PST)
Received: (from dbg@localhost) by (8.15.2/8.15.2/Submit) id v1G2JeOZ015346; Wed, 15 Feb 2017 18:19:40 -0800 (PST)
Date: Wed, 15 Feb 2017 18:19:40 -0800
From: Daniel B Giffin <>
To: "Black, David" <>
Message-ID: <>
References: <>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Disposition: inline
In-Reply-To: <>
Archived-At: <>
Cc: tcpinc <>
Subject: Re: [tcpinc] WGLC for draft-ietf-tcpinc-tcpcrypt - draft header
X-Mailman-Version: 2.1.17
Precedence: list
List-Id: "Working group mailing list for TCP Increased Security \(tcpinc\)" <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Thu, 16 Feb 2017 02:19:44 -0000

Sorry for the oversight on the document category!  I've
submitted a new draft labeled Experimental.

In the new draft (-05) I've also changed the denial-of-
service discussion slightly to make clear that there is no
special vulnerability of tcpcrypt above what TCP suffers:

   To gain middlebox compatibility, tcpcrypt does not protect TCP
   headers.  Hence, the protocol is vulnerable to denial-of-service from
   off-path attackers just as plain TCP is.  Possible attacks include
   desynchronizing the underlying TCP stream, injecting RST or FIN
   segments, and forging rekey bits.  These attacks will cause a
   tcpcrypt connection to hang or fail with an error, but not in any
   circumstance where plain TCP could continue uncorrupted.
   Implementations MUST give higher-level software a way to distinguish
   such errors from a clean end-of-stream (indicated by an authenticated
   "FINp" bit) so that applications can avoid semantic truncation

I didn't mention on-path attackers because it should be
clear that they have arbitrary DOS capability.

Thanks for the suggestion.