Re: [Iot-onboarding] some straw-man charter text for an IoT Operational Security WG

Eliot Lear <> Mon, 16 September 2019 11:26 UTC

Return-Path: <>
Received: from localhost (localhost []) by (Postfix) with ESMTP id 552B4120019; Mon, 16 Sep 2019 04:26:03 -0700 (PDT)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: -14.501
X-Spam-Status: No, score=-14.501 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIMWL_WL_MED=-0.001, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_HI=-5, SPF_PASS=-0.001, USER_IN_DEF_DKIM_WL=-7.5] autolearn=ham autolearn_force=no
Authentication-Results: (amavisd-new); dkim=pass (1024-bit key)
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id WdbpA9qSlm4F; Mon, 16 Sep 2019 04:26:00 -0700 (PDT)
Received: from ( []) (using TLSv1.2 with cipher DHE-RSA-SEED-SHA (128/128 bits)) (No client certificate requested) by (Postfix) with ESMTPS id 632B9120837; Mon, 16 Sep 2019 04:25:59 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple;;; l=28012; q=dns/txt; s=iport; t=1568633159; x=1569842759; h=from:message-id:mime-version:subject:date:in-reply-to:cc: to:references; bh=tix3KOgNYl6HFxtOAkmhBx5rgQQ/QNNOIbtwIOXwPpc=; b=FVfSEasDRbZHYLxL7e1h8TneStRqUZkP7EAv8p+tlKqEeN5eLDRU8Bcg T+h511OSKl8ULvpea4qWD+w/SOkehiNe8y9CaIDl1iPVUpXqdnoUCduWM VcQkSMHyBCUnSl+jWurc0btHMhOHZElRP9E7jcufdWJ4f8J0xlWN4n4G5 E=;
X-Files: signature.asc : 488
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: =?us-ascii?q?A0BZAABocH9d/xbLJq1gBhoBAQEBAQI?= =?us-ascii?q?BAQEBBwIBAQEBgWeBaQWBF1MgEiqEIYh8iBl+mDWBYwQCBwEBAQkDAQEYAQU?= =?us-ascii?q?RAQGBS4J0AoMSOBMCAwkBAQQBAQECAQUEbYUuDIVKAQEBAQIBAQEYCUgDCwU?= =?us-ascii?q?LCxIGFRUCAiciDgYTFAUCgwcBgXsPD6oJgUqBMh+EGAEDAgEBDw9vhGoKBoE?= =?us-ascii?q?0gVGKP4F/gREnH4FOSQcuPoJhAQECAYEZCQkBCwcBCQhNgkwygiYEf4t+Cg8?= =?us-ascii?q?DiFiWeYIsgi6BE4NEjXsbgjWHR4N+ix+KEYwEjWWDEQIEBgUCFYFpIWdxMxo?= =?us-ascii?q?IGxU7KgGCQT6CCzFvAQmCQYUUhUE+AzABAQGOHw8Xgi4BAQ?=
X-IronPort-AV: E=Sophos;i="5.64,512,1559520000"; d="asc'?scan'208,217";a="16861838"
Received: from (HELO ([]) by with ESMTP/TLS/DHE-RSA-SEED-SHA; 16 Sep 2019 11:25:56 +0000
Received: from [] ([]) by (8.15.2/8.15.2) with ESMTPS id x8GBPtCI027680 (version=TLSv1.2 cipher=DHE-RSA-AES256-GCM-SHA384 bits=256 verify=NO); Mon, 16 Sep 2019 11:25:56 GMT
From: Eliot Lear <>
Message-Id: <>
Content-Type: multipart/signed; boundary="Apple-Mail=_7B7D88EC-12BC-464B-B4F5-7FE078960488"; protocol="application/pgp-signature"; micalg=pgp-sha256
Mime-Version: 1.0 (Mac OS X Mail 12.4 \(3445.104.11\))
Date: Mon, 16 Sep 2019 13:25:54 +0200
In-Reply-To: <>
Cc: Brian E Carpenter <>, Mohit Sethi M <>, Michael Richardson <>, "" <>, "" <>
To: Mohit Sethi M <>
References: <> <> <> <> <> <>
X-Mailer: Apple Mail (2.3445.104.11)
X-Outbound-SMTP-Client:, []
Archived-At: <>
Subject: Re: [Iot-onboarding] some straw-man charter text for an IoT Operational Security WG
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: Discussion of IoT onboarding mechanisms <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Mon, 16 Sep 2019 11:26:04 -0000

> On 16 Sep 2019, at 11:19, Mohit Sethi M <> wrote:
> Hi Brian,
> A protocol that has many dependencies (or gaps) is not ideal from my perspective. A protocol that requires V from chip vendor, W from the device manufacturer, X from certificate authorities, Y from the network administrator, and Z from the user is not a good protocol.

To a point I agree, depending on what V, W, X, Y, and Z are.  If they are things that are at all painful for any player to provide, the &ing of those conditions make success highly unlikely.
> Note that this is a hypothetical protocol that thankfully doesn't exist.
> I agree with your statement that IETF should not produce standards with gaps. But the solution is not to fill those gaps. A better solution is that our protocols shouldn't have too many dependencies in the first place.

To me I think we are still in early days, all of us, on this problem. Protocols like SZTP or BRSKI or the TEAP extensions we are doing probably require a bit of reconciliation.  What is key, at least for cloud-based provisioning, is the voucher.  What is also key, it seems to me, is an interface like EST.

What we heard in our call with Randy, for instance, was that in that vertical, we have to take into account lengthy SI chains, and we have to do so without client interactions.  One way to address this from an onboarding perspective is to really drive an architecture document such that the components can be shown to fit well together, in a bit of a plug and play mode.

That discussion, it seems to me, could easily drown out all other discussions and seems quite disparate from MUD in particular, which leads me to wonder again if we are being too broad.


> --Mohit
> On 9/12/19 11:28 PM, Brian E Carpenter wrote:
>> Hi Mohit,
>>> With my limited
>>> experience of IETF, I certainly don't think IETF is in the business of
>>> building ecosystems
>> No, but the open source community that uses our standards definitely is in
>> that business, so interoperable standards need to help such work.
>>> (and neither should it be).
>> However, the IETF should not produce standards with gaps that encourage
>> proprietary ecosystems that allow customer capture. That's exactly why ANIMA
>> includes a reference model as well as specific standards. It encourages each
>> vendor to provide a MASA and encourages network operators to mix and match
>> products from multiple vendors.
>> Whether the BRSKI/MASA model generalises beyond autonomic networks remains to
>> seen, but again: it was not designed for IoT.
>> Regards
>>    Brian Carpenter
>> On 12-Sep-19 23:33, Mohit Sethi M wrote:
>>> Hi Brian,
>>> IETF is in the business of building tools (i.e. open specifications with
>>> running code) for developers. And these tools are best built in working
>>> groups which have the expertise on them. Most people outside the ANIMA
>>> community would not know what is a MASA. Similarly, most people outside
>>> the EMU community would not know that Session-Ids for fast
>>> re-authentication must be exported by all EAP methods.
>>> I agree with folks that there may be multiple solutions that are
>>> relevant to the bootstrapping problem. But each of those should
>>> developed in working groups where the relevant expertise is present. One
>>> could argue that that we would end up developing different solutions for
>>> the same problem in silos. However this why we have the IESG ,the
>>> directorates, and liaisons to other standards bodies. It ensures that we
>>> are aware of related work ongoing in different fora. With my limited
>>> experience of IETF, I certainly don't think IETF is in the business of
>>> building ecosystems (and neither should it be).
>>> --Mohit
>>> On 9/11/19 11:35 PM, Brian E Carpenter wrote:
>>>> Hi Mohit,
>>>> On 12-Sep-19 07:21, Mohit Sethi M wrote:
>>>>> Hi Michael,
>>>>> I wonder why a new working group is needed and why this work cannot be pursued in some of the existing working groups?
>>>>> I suppose ANIMA was recently re-chartered (and can be re-chartered again).
>>>> We've been very insistent that ANIMA is scoped for professionally managed networks. That is not, IMHO, a reasonable restriction for IoT; so the ANIMA scope is narrower. Also, ANIMA is scoped for autonomic management, with bootstrap and security being only part of the requirements; in that sense, the ANIMA scope is broader.
>>>>> EMU is currently going over the re-charter text.
>>>> I know little about EAP, but it seems to me that although it may well be a primary tool for on-boarding, it is only a tool, and not a complete ecosystem. The "Thinking through onboarding" thread scopes the wider problem nicely.
>>>> Regards
>>>>     Brian
>>>>> Also, you write:
>>>>>> adopt a cloud-less (MASA-less, AAA-less) onboarding mechanism (possibly a version of EAP-NOOB),
>>>>> There is clearly some misunderstanding about EAP-NOOB here. EAP-NOOB is specifically intended for registering new IoT devices on a server (and associating it with a user account). The fact that it provides network-access credentials is a bonus. Please have a look at slides 3-10 here: <>
>>>>> You clearly see a AAA server in the figures. So calling it AAA-less doesn't make sense.
>>>>> --Mohit
>>>>> On 9/4/19 10:45 AM, Michael Richardson wrote:
>>>>>> I wrote this last week, and passed it around for obvious objections.
>>>>>> <>
>>>>>> You can use the crayon/edit button on github to suggest changes, or email.
>>>>>> Charter for Working Group
>>>>>> The words "Internet of Things" or IoT have come to mean anything and
>>>>>> everything to a wide group of technology players. The IETF has been working
>>>>>> on a wide variety of protocols for use by machine to machine
>>>>>> communication. This include CoAP, CBOR, 6TISCH, ROLL, SUIT, NETCONF SZTP,
>>>>>> T2TRG, ANIMA's BRSKI onboarding protocol, and most recently RFC8520, the
>>>>>> Manufacturer Usage Description.
>>>>>> The IETF has tried to focus on categories of what limited things can do, and
>>>>>> this has resulted in a number of useful documents from the Light-Weight
>>>>>> Implementation Guide (LWIG). RFC7228 is a key product, having provided
>>>>>> terminology and scaling understanding to the entire industry. All of this has
>>>>>> been about scaling the Internet technologies to small devices and constrained
>>>>>> networks. In aggregate, these devices on small networks present a significant
>>>>>> operational risk to the Internet as a whole, and even to individual
>>>>>> Enterprise, simply due to their numbers, and lack of opportunity for regular
>>>>>> human supervision.
>>>>>> IoT devices already exist today in vast numbers. Most devices that people are
>>>>>> personally familiar with are in the BlueTooth Connected devices, or
>>>>>> Web-Connected devices that use WiFi to reach servers on the Internet ("the
>>>>>> Cloud"). Increasingly, the IETF view of machine to machine communications are
>>>>>> colinizing new greenfield situations. The IETF notion of autonomous networks
>>>>>> of devices is still a minority view compared to the market IoT industry of
>>>>>> cloud-only connected devices, but the transition is occuring.
>>>>>> RFC8520 was created to bridge the gap between devices wholly controlled by a
>>>>>> local operator (such as Enterprise IT), and devices which can not assume any
>>>>>> infrastructure at all, and must rely entirely on cloud communications for
>>>>>> command and control.
>>>>>> This working group concerns itself with Operational Security of IoT systems.
>>>>>> This includes:
>>>>>> * factory provisioning of devices
>>>>>> * onboarding of devices
>>>>>> * access control of devices to network resources
>>>>>> * administrative control of devices
>>>>>> * asset management of devices, as it pertains to software/firmware versions
>>>>>> * isolation/quarantine of devices
>>>>>> * remediation of broken devices
>>>>>> * end of life management of devices
>>>>>> The WG is chartered explicitely to work on MUD (RFC8520) and extensions to it.
>>>>>> The WG is chartered to work on onboarding protocols, specifically including
>>>>>> derivaties of BRSKI (RFC-tbd), but not limited to just that protocol.
>>>>>> The WG is not expected to pick a winner, and is encouraged to work on a
>>>>>> multitude of use-case specific protocols: better to get one use case right,
>>>>>> than to be too-complex jack of all trades.
>>>>>> The WG is expected to articulate clear applicability statements for each
>>>>>> protocol. The WG is expected to produce concise Roadmap documents that
>>>>>> explain how a variety of IETF (and other) protocols can work together to
>>>>>> satisfy the Operational needs of specific IoT areas. These roadmap documents
>>>>>> needn’t result in RFCs.
>>>>>> Neither the WG nor the IETF has exclusivity here, and an ideal document would
>>>>>> be one that the WG helps to start, but a specific industry alliance becomes
>>>>>> the lead editor for.
>>>>>> There will be coordination with many other WGs beyond the list above, and
>>>>>> this WG may accept applicability statement work from other WGs about specific
>>>>>> ways to deploy their protocols.
>>>>>> The WG will operate through a series of virtual interim meetings. This is
>>>>>> driven by a need to interact regularly with other industry grouops, and due
>>>>>> to the variety of topics which will not always be able to get quorum as a
>>>>>> committee of the whole.
>>>>>> {unusual, maybe not charter appropriate, but rather saag-like}
>>>>>> During in-person meetings, the WG will deal with typical status and document
>>>>>> progress issues during one hour (or less) of the time, and during another
>>>>>> hour, will be open to slideware presentations and tutorials on current IETF
>>>>>> or other-SDO IoT efforts. The goal of these presentations is to quickly
>>>>>> communicate current IoT systems state to the rest of the IETF.
>>>>>> It is acknowledged that part of the value is in YouTube content, and some
>>>>>> content should be done at IAB tech plenaries rather than at the WG.
>>>>>> The initial set of work items is included below as milestones, which only
>>>>>> require AD approval.
>>>>>> Milestones
>>>>>> * adopt the constrained-voucher/constrained-BRSKI work from ANIMA.
>>>>>> * adopt the dtsecurity-zero-touch work from 6tisch, which can not finish before a LAKE finishes.
>>>>>> * create a list of a series of MUD extensions, and revise this milestone
>>>>>> * adopt a cloud-less (MASA-less, AAA-less) onboarding mechanism (possibly a version of EAP-NOOB), that can be used at the retail level.
>>>>>> * negotiate with EMU WG on how to proceed with TEAP-BRSKI, and revise this milestone.
>>>>>> * adopt a cloud-driven onboarding mechanism that can be used in completely offline situations without requiring renewals (perhaps revising RFC8366).
>>>>>> ....
> --
> Iot-onboarding mailing list