Minutes IETF100: t2trg

Meeting Minutes Thing-to-Thing (t2trg) RG
Title Minutes IETF100: t2trg
State Active
Other versions plain text
Last updated 2017-12-04

Meeting Minutes

T2TRG Summary meeting
November 14, 2017
IETF 100, Singapore

Chairs: Carsten Bormann, Ari Keränen
Notes: Dirk Kutscher , Solomon Kembo

TUESDAY, November 14, 2017. 1550-1750  Afternoon Session II, Padang

Chairs: Intro, RG Status
Ari: Correct link of the agenda is on the mailing list Carsten: T2TRG is a
research group not an IETF working group. We focus on issues with opportunities
for IETF standardization. Next Meetings: WISHI Hackathon follow-up (27
November) as we couldn't finish our work during the Hackathon. NDSS Workshop
February 18 (decentralised IoT Security) submit papers by December 1. We are
planning to join hackathons of other organisations like OCF. We have one
research document ready for publication "State-of-the-Art and Challenges for
IoT Security". RESTful Design of IoT document adopted as RG item.

15:50 Chairs: Meeting reports for Berlin and OCF
Carsten: Quick report on meeting in Berlin.
RIOT Summit & ACM ICN 2017 conference
Meeting had a lot of topics but I will highlight three topics:
    1. (Ad-hoc) What is IoT?
definition of "thing": Internet node that has a foot in the physical world
the role of "constrainedness": often a property of things
scalability requiring frugality in cost, power usage etc. (scaling down)

   2. Coexistence
   We had short seminar on coexistence in Chicago.
   We are trying to understand interesting coexistence properties such as
   Spectrum and IP networks

   Also had ten other topics we cannot cover today.

Ari: We had packed agenda with 16 topics at the OCF meeting before the IETF 100
meeting. IETF and OCF will work together through monthly calls. OCF to review
Resource Directory draft.

longer list of action items for collaboration with OCF

WISHI Hackathon: 2 hour hackathon on Sunday
- semantic interop interworking
- next call: Monday, Nov 27, will be a practical experiment

Q Ali: have you contacted ITU-T?
Ari: don't think we had them at our previous meetings
Ali: At Buenos Aires, they had a lot of IoT security info
Ari: please let us know who to involve

Q Jabber: posted a link to the chat room

16:10 R. Moskowitz: Small Crypto for Small IoT
Robert: We had a lot of vendors that left because their devices were too small.
I will talk of a standard that should bring the vendors back into the room. -
Keccak is a whole new approach from traditional cryptography since CAESAR
crypto. - "sponge function" approach - learn  more from https://keccak.team -
selected for SHA-3 - is well optimized for 32-bit, 64-bit and multi-core CPUs
and large messages - Obligatory picture: what is a sponge?
https://keccak.team/files/CSF-0.1.pdf - Keccak is a complete symmetric crypto
solution (crypto hash, keyed hash, PRF, data encryption) - Single primitive to
implement replacements for AES, HMAC, SHA-2 - highly parameterized and comes in
all sizes 25, 50, 100, 400, 800, 1600 bits - really good for small messages -
defines 24 rounds which are easy to increase, which add future proofing for
unknown attacks. - truly suitable for constrained IoT - KMAC outperforms HMAC,
with smaller code size - Ketje SR outperforms AES-CCM -
https://tools.ietf.org/html/draft-moskowitz-small-crypto-00 - yet to move some
diagrams that are in pdf form into the drafts format. - Next steps:
    - add Keccak b=400 to protocol specs
    - develop PoC IoT devices
- we can have a digital certificate for constrained IoT devices using Keccak
- Am looking for people to work with me on this project
- Bob: suggesting to demonstrate industry need to NIST

Gabriel M.: This is not well known. Existing mode: crypto stuff in CFRG as
consulting body. This should be useful beyond IoT. Encouraging to talk to CFRG
to find out how they see Keccak and obtain some guidance. Bob: going to take
this to CFRG. But this is a good forum to get discussion started. Ali:
centralization or decentralization? Bob: medical example -- it's a whole
different discussion. Focus here is communication security. Ali: Today's
approach: security first. Distributing information to multiple sites and
recipients. Not limited to one realm. Bob: protecting data at rest would still
benefit from having fast, efficient ciphers. Some things are really orthogonal.
Alex P: LPWAN co-chair perspective: seems to be interesting work for LWPAN
community Bob: 802.15 study group on ng security Mohit: outperforming HMAC:
from what I have seen, the code size is not the limiting factor, but the memory
usage is (and execution time) Bob: more efficient in general (less power, using
simpler operations) from what I have seen from Keccak documents

16:35 Xavier de Foy: IoT Edge Computing Survey and Gap Analysis

Xavier presenting
- light-weight and open edge computing -- has different requirements from
existing (telco-driven) edge computing - 5G-CORAL research that combines telcom
edge computing and fog - Intelligent IoT Gateway Model Products and
Projects(Bosch, Siemens, Microsoft, Amazon, EdgeX Foundry) - Typical
Protocols(HTTPS, MQTT, AMQP, COAP, OPC UA, DDS) - OpenFog architecture linked
to IEEE P1934 WG - emerging trends: data-oriented networking and in-network
computation, for example Information-Centric Networking (ICN) and Named
Function Networking (NFN) - Gap analysis:
    - IoT EC requires distributed computing model
    - requiring a more open model
    - IoT EC should support lower end devices which may introduce additional
    technical challenges
- Next Steps

    Dirk: Booked the Butterworth room from 15:30 to 17:00  tomorrow (Wednesday)
    for a side meeting. [Notes for that meeting are now at: ] Erik: NAT
    Traversal could be an issue (engineering question) Erik: DINRG is looking
    at decentralized network infrastructure which seems to be related

    Ari: chair-chat off, I am also very much interested in the NAT traversal
    work; we already have the Thin ICE activity and could have a look if that
    solves the problems and/or we need something else

Liang GENG: Problem Statement of Edge Computing beyond Access Network for
Industrial IoT Slides:

Mike McBride presenting
- we narrow focus of Edge computing with special emphasis on the factory setting
- we had side meeting at last IETF meeting and we have created a problem
statement draft - identified first set of requirements - Beyond Edge Computing:
distribute as much as you can - Further discussion: Thursday 20:00 -- 21:30

Ali: you are talking about Industrial area: are you talking about how they are
going to migrate from Bacnet to IP? Mike: Best answer would be to work together
in collaboration with vendors Dirk: Huge conversion currently going on
home-grown proprietary solutions; there are going to be security and interop
problems. Erik: specific what virtualization technologies? Mike: Depends but
specific application requirements Eve Schooler: Gap analysis is great. Many of
them are focused on frameworks and architectures. IETF is good at looking
specific protocols and interfaces for these architectures. Looking at ecosystem
of compomenents, where are the places where we can design protocols, e.g.,
federation, routing, in-network caching. Edge/Fog computing shifts discussion
away from only constrained devices. It might be worthwhile for T2TRG to
consider non-constrained devices as well. Dave Thaler: protocol gateways are
often used for translating between field buses and IP etc. * Matthias: at WoT
have quite a lot of the features talked here. WoT tries to take IETF protocols
and use them e.g. in industrial environment and trying to harmonize the
environment. Good to have look at that. IG that does exploration and WG that
does standards work.

17:10 Michael McCool (remote): WISHI semantic interop of AVS and IoT

- evolution of IoT towards "ambient" computing
- different levels of interoperability: semantic, structural, syntactic
- IoT Ontology: iotschema.org (Things as capability bundles)
- W3C Web of Things Architecture: thing description (TD), Scripting API,
Binding API: does not require special gateways - PoC development, testing
several different approaches

17:40 [Chairs: Meeting Planning, Wrapup -- already done at intro]

Session ended at 17:50