IETF
t2trg@jabber.ietf.org
Tuesday, November 14, 2017< ^ >
Dave Thaler has set the subject to: https://github.com/t2trg/2017-07-wishi
Room Configuration
Room Occupants

GMT+0
[07:38:55] meetecho joins the room
[07:45:12] HU TUO joins the room
[07:45:12] Michael McCool joins the room
[07:45:12] Lorenzo Miniero joins the room
[07:46:38] Ann Krieger joins the room
[07:47:50] Lihao Chen joins the room
[07:50:52] Yoshiro Yoneya joins the room
[07:54:32] Michael Koster joins the room
[07:55:29] Michael Breuer joins the room
[07:55:38] <Michael McCool> ari, just sent you updated slides by email; please use if possible, but if not, the old versions are ok
[08:01:05] <Michael McCool> WoT also has regular plugfests
[08:02:15] limyinglu joins the room
[08:02:54] <Michael McCool> should address concept of "actions" in RG Doc...
[08:05:50] <Michael McCool> this definition is somewhat limited and unnecessary: "IoT devices" are part of a system, which will also include mobile phones, cloud services, gateways, fog compute, etc.
[08:06:17] <Michael McCool> IoT devices as defined here are just "leaf sensors" in such a system
[08:07:34] <Michael McCool> security... more important in the "real world"
[08:07:45] <Michael McCool> eg DDoS
[08:09:03] <Yoshiro Yoneya> Please speak more loudly or close to mic.
[08:11:38] Dave Thaler joins the room
[08:11:41] <Michael McCool> ARI: my presentation: https://github.com/w3c/wot/blob/master/PRESENTATIONS/IoT%2C%20Semantic%20Interop%2C%20and%20Voice%20Control%20-%20McCool.pdf
[08:18:24] <Michael McCool> http://www.itu.int/en/ITU-T/gsi/iot/Pages/default.aspx
[08:18:56] Barry Leiba joins the room
[08:19:14] g.e.montenegro joins the room
[08:19:27] <Michael McCool> scribing...
[08:19:49] <Michael McCool> comment about engaging with ITU especially on security
[08:20:01] <Michael McCool> talk: Robery Moskowitz
[08:20:01] Behcet Sarikaya joins the room
[08:20:15] <Michael McCool> issue: many device vendors have left; devices "too small"
[08:20:29] <Michael McCool> want to work on standards that scale down better to small devices
[08:20:40] <Michael McCool> and also improve efficiency on larger devices
[08:20:42] Dave Thaler has set the subject to: https://github.com/t2trg/2017-ietf100
[08:21:24] <Michael McCool> Keccak: new approach to symmetric cryptography
[08:21:45] <Michael McCool> complete break, based on new concept: sponge function
[08:22:14] Behcet Sarikaya leaves the room
[08:22:23] <Michael McCool> at maximum strength, was selected for SHA-3
[08:22:50] <Michael McCool> basis for Ketje cipher in CAESAR competition
[08:22:59] <Michael McCool> also FIPS 202
[08:23:06] Behcet Sarikaya joins the room
[08:23:08] <Michael McCool> ... what is a sponge?
[08:23:14] <Michael McCool> absorbs and squeezes
[08:23:56] <Michael McCool> https://keccak.team/files/CSF-0.1.pdf
[08:24:16] <Michael McCool> gives complete solution
[08:24:35] akeranen joins the room
[08:24:38] <Michael McCool> hash, keyed hash, pseudo-random function, data encryption, authentication
[08:24:47] <Michael McCool> all major necessary primitives
[08:25:07] <Michael McCool> so one hardware primitive (or software library) supplies all functions
[08:25:21] <Michael McCool> also parameterizable, can vary strength easily
[08:25:39] <Michael McCool> defines a bitrate, rather than a blocksize
[08:25:53] <Michael McCool> three variables
[08:25:56] <Michael McCool> r, b, c
[08:26:02] <Michael McCool> r is bitrate
[08:26:11] <Michael McCool> c is capacity
[08:26:15] <Michael McCool> r = b - c
[08:27:12] <Michael McCool> there is also another parameter: number of rounds
[08:27:24] <Michael McCool> currently uses 24; known attacks against 6
[08:28:15] akeranen leaves the room
[08:28:34] akeranen joins the room
[08:28:54] Dirk Kutscher joins the room
[08:29:25] <Michael McCool> b = 400 good for small devices
[08:30:52] HU TUO leaves the room
[08:35:36] <Michael McCool> discussoin
[08:35:38] <Dave Thaler> gabriel montenegro at mic
[08:35:47] <Michael McCool> this is not very well known
[08:35:54] <Michael McCool> any reason why this should not ...
[08:35:59] <Michael McCool> we have the cfrg
[08:36:04] <Michael McCool> that is a consulting body
[08:36:13] <Michael McCool> they can look at this and come out with an opinion
[08:36:25] <Michael McCool> is good, useful for other things in addition to ioT
[08:36:33] <Michael McCool> would be good to get their opinion
[08:36:44] <Michael McCool> then other groups can take that guidance
[08:36:51] <Michael McCool> would that model not apply?
[08:37:00] <Michael McCool> speaker: yes, will be taking it to the crg
[08:37:26] <Michael McCool> rather than just saying "here is another cipher" want to state "where the interest lies"
[08:37:38] <Michael McCool> consider the low-resource router
[08:37:53] <Michael McCool> here is a way you can protect the IP flow on smaller devices
[08:38:00] <Michael McCool> so want to start with this audience...
[08:38:05] <Michael McCool> next question:
[08:38:23] <Michael McCool> what is your application?   centralization or decentralization?
[08:38:23] <Dave Thaler> don't know mic speaker's name
[08:38:42] <Michael McCool> where do you take this wrt the topology?
[08:38:49] <akeranen> Ali
[08:38:53] <Michael McCool> speaker: sensors start data collection chain
[08:39:01] <Dave Thaler> yeah Ali thanks Ari
[08:39:04] <Michael McCool> back to the hospital, gets into big data
[08:39:25] <Michael McCool> having something like this for sensor end is necessary
[08:39:31] Behcet Sarikaya leaves the room
[08:39:36] <Michael McCool> there is a whole other discussion around big data
[08:39:47] <Michael McCool> not about how data at rest is encrypted
[08:40:07] <Michael McCool> question: but is part of entire system design
[08:40:13] <Michael McCool> security comes in first
[08:40:24] <Michael McCool> which is why I asked about centralization or decentralized
[08:40:47] <Michael McCool> right now, multiple sites, practitioners, hospitals
[08:41:16] <Michael McCool> speaker: if I want to protect data at rest efficiently, then yes, might want to select light and fast encryption
[08:41:24] <Michael McCool> if you want to have the big data discussion, we can
[08:41:31] <Michael McCool> but the concerns are orthogonal
[08:41:44] <Dave Thaler> alexander pelov at mic
[08:41:50] <Michael McCool> questioner: I like your approach, but want to know how to drive
[08:42:09] <Michael McCool> new questioner: I was not aware of this, thanks
[08:42:14] akeranen leaves the room
[08:42:18] <Michael McCool> how to move forward
[08:42:27] <Michael McCool> speaker: there is new study group
[08:42:29] Mert Ocak joins the room
[08:42:38] <Michael McCool> people usually talk about "bigger" security
[08:42:48] <Michael McCool> but will use it as an excuse to get smaller stuff in too
[08:42:57] <Michael McCool> new question: performance
[08:43:12] <Dave Thaler> Mohit Sethi at mic
[08:43:14] <Michael McCool> but from what I have seen, code size is not the most limiting factor
[08:43:21] <Michael McCool> how does it compare in time and memory use?
[08:43:28] <Michael McCool> speaker: I have seen those numbers
[08:43:34] <Michael McCool> all are lower across the board
[08:43:42] <Michael McCool> but I will have to go pull those out
[08:43:57] <Michael McCool> it does take less power, using simpler operations
[08:44:04] <Michael McCool> from what I have seen in documents
[08:44:36] <Michael McCool> speaker: anything else? if not, thanks for your time, let's figure out together how to move this forward
[08:44:40] <Michael McCool> next topic
[08:44:46] <Michael McCool> edge computing
[08:44:53] <Michael McCool> survey and gap analysis
[08:45:55] <Michael McCool> Dirk and Eve
[08:46:24] limyinglu leaves the room: Stream reset by peer
[08:46:32] Lihao Chen leaves the room
[08:47:24] <Michael McCool> (see slides and speaker notes)
[08:50:09] <Mert Ocak> It seems that it is not possible to join the video stream, Meetecho rejects as "too late". However, other wg meetings going on at the moment allows participation. Maybe a bug or misconfiguration?
[08:50:29] <meetecho> Mert Ocak: looking into that
[08:51:16] limyinglu joins the room
[08:51:38] <meetecho> It should be fixed now, sorry about that
[08:53:31] Mert Ocak joins the room
[08:54:16] Lihao Chen joins the room
[08:54:34] Mert Ocak leaves the room
[08:54:37] Mingliang Pei joins the room
[08:55:31] Mert Ocak_3698 joins the room
[08:55:57] Mert Ocak leaves the room
[08:56:15] Mert Ocak_3698 leaves the room
[08:56:26] Mert Ocak joins the room
[09:02:05] <Dave Thaler> Dirk Kutcher at mic
[09:02:12] <Dave Thaler> Kitscher
[09:02:13] <Michael McCool> dirk: there is a followup meeting
[09:02:16] <Dave Thaler> Kutscher
[09:02:17] <Michael McCool> on edge computing
[09:02:28] <Michael McCool> various telco-motivated approaches
[09:02:41] <Michael McCool> typically extending cloud compting to the edge
[09:02:51] <Michael McCool> in a t2t environment with constrained nodes
[09:02:59] <Michael McCool> might have different requirements
[09:03:16] <Michael McCool> if interested in discussing more, have book breakout room tomorrow; please come talk to me
[09:03:21] <Dave Thaler> Erk Nordmark at mic
[09:03:31] <Michael McCool> quite useful to have this overview
[09:03:34] <Michael McCool> practical consideration
[09:03:38] <Michael McCool> NAT traversal
[09:03:49] <Michael McCool> East-west communicatoin at the edge
[09:03:54] <Michael McCool> bump into NAT
[09:04:11] <Michael McCool> is a practical problem, should not lose sight of it
[09:04:12] <Michael McCool> also
[09:04:18] <Michael McCool> can edge be more autonomous
[09:04:31] <Michael McCool> how do we handle security, etc?  definitely need research here
[09:04:45] <Michael McCool> ari: have been looking at NAT
[09:04:51] <Michael McCool> interested in whole problem space
[09:05:12] <Michael McCool> another short 10-minute presentation on edge computing
[09:05:48] <Michael McCool> Mike McBride
[09:05:55] <Michael McCool> more narrow focus of edge computing
[09:06:02] <Michael McCool> huawei
[09:06:07] <Michael McCool> industrial internet forums
[09:06:13] <Michael McCool> focus is on industrial
[09:06:19] <Michael McCool> (see slides)
[09:11:59] <Michael McCool> BEC: Beyond Edge Computing
[09:13:46] <Michael McCool> use case: connected elevator
[09:14:02] <Michael McCool> at openlab in nanjing
[09:15:31] <Michael McCool> questions....
[09:15:37] <Michael McCool> Ali
[09:16:01] <Michael McCool> industrial area... how to migrate from existing system, eg BACnet
[09:16:12] <Michael McCool> very difficult to migrate from BACnet to IP
[09:16:32] <Michael McCool> a: broader question; one way is vendor partnerships
[09:16:51] <Michael McCool> already happening...
[09:16:55] <Michael McCool> (who is speaking?)
[09:17:01] <Michael McCool> this is relevant
[09:17:02] <Dave Thaler> Dirk
[09:17:18] <Michael McCool> companies are already using their own proprietary approaches
[09:17:28] <Michael McCool> is going to be a security and interop problem in the future
[09:17:34] <Michael McCool> new speaker...
[09:17:57] <Michael McCool> virtualization... what specifically are you looking at?  networking, or also compute, etc?
[09:18:12] <Michael McCool> a: in each industry are certain apps needed and used
[09:18:30] <Michael McCool> for each vertical, need specific approaches to virtualize
[09:18:42] <Michael McCool> q: such as breaking it up in a virtualized way?
[09:18:44] <Michael McCool> a: yes..
[09:18:46] <Michael McCool> Eve:
[09:18:56] <Michael McCool> ... been part of dialog here on edge computing
[09:19:07] <Michael McCool> wanted to emphasize... love the gap analysis
[09:19:13] <Michael McCool> are focused on frameworks, etc.
[09:19:17] <Michael McCool> IETF does well
[09:19:22] <Michael McCool> is spec of comm protocols
[09:19:31] <Michael McCool> seems to be missing from some of these conversations
[09:19:33] <Michael McCool> quesion is
[09:19:38] <Michael McCool> looking at this ecosystem
[09:19:50] <Michael McCool> where are the places we can design communication protocols?
[09:19:54] <Michael McCool> gap analysis...
[09:20:24] <Michael McCool> can you identify and prioritize gaps having to do with comm protocols
[09:20:27] <Michael McCool> next topic
[09:20:33] <Michael McCool> what is interesting about edge
[09:20:45] <Michael McCool> is that it shifts dicussion away from constrained devices
[09:21:09] <Michael McCool> so... what kinds of functionality would be useful for IoT that don't necessarily have to run on constriained devices?
[09:21:20] <Michael McCool> a: yes, we intend to follow up with a gap analysis
[09:21:23] <Michael McCool> Dave T:
[09:21:29] <Michael McCool> comment on BACnet
[09:21:35] <Michael McCool> slide had protocol gateway
[09:21:38] <Michael McCool> typical approach
[09:21:47] <Michael McCool> BACnet is not the only such case
[09:21:54] <Michael McCool> Matthias K:
[09:21:59] <Michael McCool> trying to understand starting point
[09:22:09] <Michael McCool> saw some parallel with WoT
[09:22:17] <Michael McCool> we are looking at similar things
[09:22:49] <Michael McCool> are trying to harmonize the gateway approach
[09:22:54] <Michael McCool> a: what WG again?
[09:23:04] <Michael McCool> q: W3C Web of Things WG
[09:23:40] Barry Leiba leaves the room
[09:24:12] <Dave Thaler> https://github.com/w3c/wot/blob/master/PRESENTATIONS/IoT%2C%20Semantic%20Interop%2C%20and%20Voice%20Control%20-%20McCool.pdf
[09:30:07] Barry Leiba joins the room
[09:31:39] Barry Leiba leaves the room
[09:37:01] Michael Breuer leaves the room: offline
[09:37:38] Mingliang Pei leaves the room
[09:37:39] g.e.montenegro leaves the room
[09:41:46] Yoshiro Yoneya leaves the room
[09:42:31] Yoshiro Yoneya joins the room
[09:46:38] Lihao Chen leaves the room
[09:51:17] Dirk Kutscher leaves the room
[09:51:22] Mert Ocak leaves the room
[09:51:45] Michael McCool leaves the room
[09:51:53] Ann Krieger leaves the room
[09:51:53] Michael Koster leaves the room
[09:51:53] Lorenzo Miniero leaves the room
[09:52:53] limyinglu leaves the room
[09:58:34] meetecho leaves the room
[10:03:46] Dave Thaler leaves the room
[10:11:11] Dirk Kutscher joins the room
[10:21:05] Dirk Kutscher leaves the room
[10:23:16] Dirk Kutscher joins the room
[10:58:52] Dirk Kutscher leaves the room
[11:15:17] Yoshiro Yoneya leaves the room
[13:29:29] Barry Leiba joins the room
[14:37:16] Barry Leiba leaves the room
[17:04:33] Yoshiro Yoneya joins the room
Powered by ejabberd - robust, scalable and extensible XMPP server Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!