20 Sep 2017

See also: IRC log


Zoltan_Kis, Kaz_Ashimura, Michael_McCool, Dave_Raggett, Kazuo_Kajimoto, Michael_Koster, Niklas_Widell, Ryuichi_Matsukura, Takeshi_Yamada, Taki_Kamiya, Tomoaki_Mizushima, Daniel_Peintner, Soumya_Kanti_Datta, Barry_Leiba, DarkoAnicic, Kunihiko_Toumura, Masato_Ohura, Sebastian_Kaebisch, Wonsuk_Lee, Keiichi_Tokuyama, Toru_Kawaguchi
Matthias, Uday
McCool, Kajimoto


<McCool> waiting until 5m past; any new agenda items, let me know

<scribe> scribenick: dsr

Quick updates

Michael reports the news that the 3 FPWDs were successfully published as planned.

<kaz> https://lists.w3.org/Archives/Public/public-wot-wg/2017Sep/0006.html

He has been working on the security workshop plans and wants to have the security document published in good time to support the workshop

<kaz> Michael McCool's message on the IEEE S&P Workshop (Member-only)

There is some 12 hours left to make changes to the workshop proposal so please send comments

PlugFest planning

Michael: any updates on the plugfest planning?

<kaz> PlugFest wiki

Matsukura-san emailed updated slides just before today’s call

<McCool> see issue #346

It is titled “Issues for TPAC Plugfest”. He presents it using the webex shared screen facility

<kaz> Matsukura-san's slides

The main focus is on communication between servients

<McCool> https://github.com/w3c/wot/issues/346 Issue for "[Plugfest] Planning for next step"

Matsukura-san talks us through each slide in turn

A device can act as a servient, or it could involve a gateway to expose a thing for another servient to consume

One issue is that the URL can be different on either side of a NAT firewall

We need some joint discussion on discovery, exchange of thing descriptions and their management.

McCool: we should support the use of ACE where practical. For events, I would like to support the subscription model, but to allow for other possibilities. I hope the approach here is descriptive and will allow for other approaches in future

kaz: the question is how to map addresses across NAT, right?

<Zakim> kaz, you wanted to ask if Matsukura-san's point is how to handle local device's address from the remote side within TD and how to transfer TD for that purpose

Matsukura-san: yes

kaz: you should make that clearer

sebastian: you showed one servient as a gateway, what do you mean by a gateway?

if you want to provide access to devices from outside of the firewall, the gateway can help

is the idea that the clients can get access direct to local devices or what is your intention?

if the former, a better term might be a bridge

Matsukura-san: the gateway provides a means to create shadows of things

<mjkoster> There is a proxy device in the gateway

(aka digital twins)

<kaz> dsr: came across the issue on URL with local/remote networks

<kaz> ... my solution was embedding the address of the gateway into the TD

<kaz> ... thousands of thermostats

<kaz> ... looks very straight-forward

<kaz> ... the question is what would be the best solution

Dave: what’s the best means for us to share ideas, as I have been working on the same problem for gateways and cloud hubs

McCool: is the issue tracker the right place for this?

kaz: the gateway can manage multiple devices on the same local area network or in the local vicinity

<inserted> ... Matsukura-san and Nimura-san would like to clarify the need/requirement for that kind of Management Thing as well

McCool: are the servients going to come from different people?

<inserted> kaz: right, so Matsukura-san would like you as well to provide one of the servients :)

McCool: the OCF bridge might fit into this pattern


What is the URL for the slides?

<kaz> Yamada-san's slides

Yamada-san: I want to talk about a proposal for eventing

we would like to have more discussion on eventing

slide 1: several times of protocols for event delivery, e.g. HTTP long poilling, server-sent events, websockets

The table shows some pros and cons for each protocol

Slide 2 shows protocol message exchanges using HTTP + WebSocket

HTTP is used to subscribe/unsubscribe and WebSockets for event delivery

Multiple devices can share the same WebSocket connection, so the message includes the device URI

The event message should include the timestamp as proposed by Mozilla

Yamada-san presents some detailed examples

questions or comments?

McCool: there is more than one way to handle events, so the TD should state which applies

The interaction model depends on the underlying protocol. e.g. HTTP vs CoAP vs WebSockets

We should assume that there are multiple ways to support eventing, and the consumer needs to know what the producer is using

sebastian: the TD needs to cover the communication metadata for that

<kaz> dsr: comment first

<kaz> ... in discussion with Ben

<kaz> ... websocket would be a good way to go

<kaz> ... using HTTPS for security and authentication

<kaz> ... why did you choose websocket just for events?

Dave: why did you choose to use WebSockets just for events and not for properties and actions? I have done that for my design.

<Zakim> dsr, you wanted to ask why not use WebSockets for properties and actions

McCool: we should discuss how to add the communications metadata for eventing

sebastian: this Friday in the TD call we will discuss metadata for eventing, could Yamada-san join that call?

Yamada-san: yes

<Zakim> kaz, you wanted to mention VW's VIWI submission and Automotive WG's VISS

kaz: I wanted to mention VW's VIWI submission and Automotive WG's VISS

<kaz> https://www.w3.org/TR/vehicle-information-service/

these are similar approaches using WebSockets for messaging

<kaz> https://www.w3.org/Submission/2016/SUBM-viwi-protocol-20161213/

we should take a look at both of these approaches

kajimoto-san: Mozilla are also interested in using WebSockets for messaging

<inserted> kaz: as discussed during the Chairs call last week, I can send a reminder to Benjamin about possible joint work

sebastian: this is not just a TD topic and is also relevant to the protocol binding discussion

Task Force reports

sebastian: last week we had a TD call with invited experts to talk to us about the SSN ontology

<kaz> TD minutes (Member-only)

we discussed TD types, we need to be careful with naming in the linked data form

this Friday we will focus on eventing and the plugfest

<zkis> https://www.w3.org/2017/09/18-wot-minutes.html (Member-only)

Zoltan talks about the scripting task force call.

We had the FPWD and discussed a small reorganisation of the repo and added Johannes as an editor

The node-wot project is considering adding the W3C license

once that happens, other people will be able to contribute to the project

<Zakim> kaz, you wanted to mention there is discussion within the Chairs group and the W3C Team about the license question

kaz: discussion within the Chairs group and the W3C Team about the license question re node-wot

this is to be clarified

Daniel: I understood from Wendy that the W3C license is equal to the MIT license, so changing should be easy

kaz: not 100% equal, we need to look at the details and get back with some recommendations

Daniel: I am waiting on this before restructuring the node-wot repo

McCool: no binding call this week, Michael Koster has some slides to present.

<kaz> https://lists.w3.org/Archives/Member/member-wot-wg/2017Sep/0024.html Koster's message

Can we defer this to next week’s call?

Michael_Koster: I could instead present at this Friday’s TD call

McCool: some pending changes on the security draft and aiming for FPWD on November 1st
... end of meeting …

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.152 (CVS log)
$Date: 2017/09/20 13:27:43 $