Re: [TLS] Middlebox Security Protocol initial drafts

Eric Rescorla <> Fri, 13 April 2018 16:26 UTC

Return-Path: <>
Received: from localhost (localhost []) by (Postfix) with ESMTP id E83C9127333 for <>; Fri, 13 Apr 2018 09:26:50 -0700 (PDT)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: -2.609
X-Spam-Status: No, score=-2.609 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_LOW=-0.7, T_DKIMWL_WL_MED=-0.01] autolearn=ham autolearn_force=no
Authentication-Results: (amavisd-new); dkim=pass (2048-bit key)
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id ftQ46kxJKW9T for <>; Fri, 13 Apr 2018 09:26:48 -0700 (PDT)
Received: from ( [IPv6:2607:f8b0:4003:c06::22b]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by (Postfix) with ESMTPS id 2256F1201FA for <>; Fri, 13 Apr 2018 09:26:48 -0700 (PDT)
Received: by with SMTP id t16-v6so8896517oih.3 for <>; Fri, 13 Apr 2018 09:26:48 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed;; s=20150623; h=mime-version:in-reply-to:references:from:date:message-id:subject:to :cc; bh=68xi8XMThHzSwwbYGPUnmdsKuzWkCZSS1McPiQFspsc=; b=CCbRQH+/olI+18sTb33BaBd5PId01ZsHjSVMB8cNC8I6AClyjR/QTqJmqDcmHuA3N7 uVIU8wzAZPjTbdB6Z3JnIun9mYNQfy7ztvnosZR0W6yBDoSMGD7AKWMHNaEe2t6ysGmx Ld8s5AFK/lng1jch7f/OKEEGQ21dv5ib6zkrM96SkpJA7kCb43I+Yepu72+H/w56WuvL WJahI7VTqfKqa/gLnIGY6ybd7Vc8aVjnWpQv7FURta7SxjOreECwWJeZJ/Xu32PwDvQq Jrv6zb5t9Ck705QoeejiIhNrI3b/UN4QTR64dT8l9wfC15ufquu3SQSJ+7QAlEomltYq Kmdg==
X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed;; s=20161025; h=x-gm-message-state:mime-version:in-reply-to:references:from:date :message-id:subject:to:cc; bh=68xi8XMThHzSwwbYGPUnmdsKuzWkCZSS1McPiQFspsc=; b=ftQ+ORL+T8Qf3zamVnjwh0grJXkYNlNROAWQUKN+v//G5xgQPPDgefmU+2Dkway607 kK1WJeVAue687l0laLjrkPai+JsHhSb44rDtBCj35SaUT+BTmNc9mJKrz9q8bAKv2tco ZAXnx0QEHzdtzwb9WHIoh6E1gAsI9euQbrJ7+rrAVyw+8vSXJzg6NT/5JHUzf1dULq+E HIZR3RQlzYCcagvmt1veji0zQa7NLAir5HuOdROBM22y3ejFaNbg2vpIe7hUu+IFEYFC 1Npyd5/IuZelgk9MvKofMHXCN86ruSbBjA7fLDYvo+8RhDNXw+OhUKUEoGs5FjHuYtiy Qo+g==
X-Gm-Message-State: ALQs6tDUzSmPNDgiC/CNWbhaEAB2Xu96mGQB0WJ26c7/9rfpIZlxmRPA O25dfATd/GT5MioXVyKMI1sXQRKp6B4U95pogYKu5Q==
X-Google-Smtp-Source: AIpwx4+AsE1HsMRvQE6hKOHwVAolKqhvOnsIVXyLmiowpyTg3E9/qTHZNjQIrb917T3qgBu9B4RT09kRf4NhYlE2rvU=
X-Received: by 2002:aca:cc53:: with SMTP id c80-v6mr9030003oig.174.1523636807509; Fri, 13 Apr 2018 09:26:47 -0700 (PDT)
MIME-Version: 1.0
Received: by with HTTP; Fri, 13 Apr 2018 09:26:07 -0700 (PDT)
In-Reply-To: <>
References: <> <>
From: Eric Rescorla <>
Date: Fri, 13 Apr 2018 09:26:07 -0700
Message-ID: <>
To: Tony Rutkowski <>
Cc: "<>" <>, Benjamin Kaduk <>,
Content-Type: multipart/alternative; boundary="000000000000adb2050569bd56d4"
Archived-At: <>
Subject: Re: [TLS] Middlebox Security Protocol initial drafts
X-Mailman-Version: 2.1.22
Precedence: list
List-Id: "This is the mailing list for the Transport Layer Security working group of the IETF." <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Fri, 13 Apr 2018 16:26:51 -0000

n Fri, Apr 13, 2018 at 9:19 AM, Tony Rutkowski <> wrote:

> Good observation. When the work started, 1.3 was a work in progress and
> the rapporteurs wanted to move forward with an initial test of concept
> based on considerable published work out there.  In addition, the use of
> 1.2 will be around for many years if not forever in some environments.

I certainly agree that this is true, but given that you're designing a new
protocol, it's not clear to me that the installed base is that relevant.

I also don't think this addresses the question of algorithms, given that
TLS 1.2 supports modern algorithms


The idea is to start a momentum for enabling middlebox visibility and
> traffic observability (Part 1), and then rolling out and encouraging
> multiple profiles.
> To give credit here, the concepts are those of pioneers like David Naylor
> and Justine Sherry. (Her slogan that "middleboxes are the drama queens of
> networking" has become legend.)
> Tony
> On Apr 13, 2018 11:36 AM, Eric Rescorla <> wrote:
> Hi Tony,
> Thanks for forwarding these.
> I haven't had time to give them a thorough review, but on a quick skim I
> notice that this seems to be based on TLS 1.2 and to use a bunch of
> algorithms we are trying to deprecate (e.g., CBC). Is there a reason not to
> start with TLS 1.3 and more modern algorithms?
> -Ekr
> On Fri, Apr 13, 2018 at 8:07 AM, Tony Rutkowski <
>> wrote:
> As everyone is aware, the decision was taken
> within the IETF TLS community not to undertake
> work perceived to compromise TLS 1.3 and related
> protocols.  At the same time, there has been
> recognition that a considerable number of use
> cases exist where alternative transport, network
> and application layer implementations are
> necessary in both network infrastructures and data
> centres.
> The work on those alternative implementations has
> proceeded over a two year period in ETSI's Cyber
> Security Technical Committee (TC CYBER). In the
> spirit of some level of collaboration between ETSI
> and the IETF community, as well as public
> transparency, CYBER is making available two draft
> parts of a Technical Specification called the
> Middlebox Security Protocol.  The motivation is to
> address one of the most difficult security
> challenges today: how to enable network operators
> and end-users to cooperate in managing encryption
> security for their applications.  The drafts are
> available at:
> Note that there are two problems being pursued
> here.  The Middlebox Security Protocol enables the
> existence of a “smart proxy” where end-users can
> be potentially aware of a middlebox in their
> traffic stream (visibility) and control what that
> middlebox sees for different purposes
> (observability). The result allows for balancing
> privacy, network operations, and security for
> different applications. With the Protocol, both
> users and providers gain the ability to grant or
> restrict the permissions for visibility and
> observability.
> Part 1 of the Middlebox Security Protocol
> specification defines the generic capabilities and
> security requirements. Additional parts define
> specific implementations in the form of profiles
> for different use cases that can be mapped to the
> Part 1 requirements.
> Part 2 provides a common profile for widespread
> network use known in the research community as
> mcTLS.  Included with Part 2 are a patch for a
> known vulnerability as well as an exemplar of use
> by Mobile Network Operators. Other profiles will
> be released over the coming months – especially
> one for data centre access control to meet the
> critical needs of enterprise network communities.
> These initial two draft specifications are
> relatively complete and stable, and derived from
> best-of-breed solutions drawn from extensive
> surveys and evaluation of the considerable
> published technical literature. However, this
> work is new, complex, and unique.
> In addition, TC CYBER is proactively sending the
> drafts to other industry standards bodies as well
> as holding a Hot Middlebox Workshop (12 June 2018)
> and Hackathon (12-13 June 2018), in Sophia-Antipolis
> France, where the coding community can seek to
> implement and hack a test implementation of Part 2.
> It is all open, and free.
> In addition to collaboration and comments on this list,
> comments are also solicited at
> --tony r