<inserted> scribenick: kaz
(Elin, Kheng, Philip from GovTech + Christine)
<jennifer> Elin Tan
<jennifer> She is an Assistant Director in Policy and Standards of the GovTech Business Strategy area
(everybody introduces themselves)
<scribe> scribenick: jennifer
McCool: TF description for the Marketing page to mention
Lagally: Approval of last minutes for 12 Jan 2021
<kaz> Jan-12
Lagally: Note to jennifer to let
colleagues to know the structure of the system
... WoT and Linked Building Data call, no representatives
here
... Look into ITU-T use cases
... WoT + OPC-UA topic to give Sebastian 15 minutes
... Currently looking at alignment with ITU-T
<mlagally> https://github.com/w3c/wot-usecases/blob/master/CONTRIBUTIONS/ITU-T-Use-case-summary.md
McCool: Identified 5 documents in ITU-T
relevant to WoT
... First for directly relate to WoT
... First four
... Description on Y.4415 is missing.
... The missing one is home network use case
... WoT Broker is a gateway service that translate between backends
and frontends
... Concept of of ITU-T is that WoT is http and there is a broker
that translates between native and http
... We need to bring that into alignment
... There is a concept of metadata we need to align with Thing
Description
Kaz: Note Y.2070 is a bit old, so maybe a bit different from the current SG 20 work. of course we can consider that document as well, though.
McCool: This is an older document that
pre-dates the W3C work
... Be aware that some of this work is legacy
... This is historical and future work is to bring it into
alignment
... There is a discovery mechanism
... The initial ITU-T WoT didn't include Discovery, but the new one
will
... The ITU-T WoT thinks of WoT services as distinct from WoT
devices
... In W3C WoT, everything is a Thing
... ITU-T has a 'web dashboard' implication
Lagally: This looks very much like a managed deployment
McCool: Which implies a broker
... Centralization means different levels
... I don't think it means centralized worldwide
... It is a centralized architecture in the idea of a broker
... ITU-T has the idea of metadata that is discoverable
... Which seems to resemble a Thing Description
... It's ambiguous if a service is also a device
... I will say this reminds me a lot of the EdgeX
architecture
... The devices themselves are reflected as services
... There is a management API to stop and start services
... Right now WoT only lets you register metadata about
services
... no way to start and stop it
... In theory this is possible with a broker
... There is confusion in some services like WoT services vs Web
services
... Confusion in some terminology
... There are services that are not correlated with any physical
device, like a dashboard
... The assumption by web means HTTP
... This is very close to what Mozilla has built
... However, our current architecture involves talking directly to
devices
... Our current architecture extends this
... We have servients, but we don't talk about a centralized thing
that manages multiple interfaces
Lagally: Is that something we could describe with a proxy concept?
McCool: It's more than a proxy
... It's a more centralized concept, we should talk about this
offline
... Web of things service architecture
... Which goes into more about services like discovery and mashups
and such
... Service profiles
... Web service description language - WSDL
... Supplanted by Open API
WSDL (Web Service Description Language)
McCool: Thing Description includes
protocol bindings
... ITU-T has a concept of Geolocation and QoS, which we don't
include
Lagally: Functional KPIs and such might want to be included
McCool: Last year in the workshop,
Matthias had a presentation on how to use QoS in a semantic
way
... The other way that QoS shows up, is that metadata shows
up
... Include things such as a last update if the device is last
alive or not
... Matthias had a proposal but didn't make an official
submission
... Captured additional references
... Multiple brokers in a federation
... We have geolocation
... Time is something in the works
... We our in the process of looking into QoS
... Time series data is pretty important
... We also need to look at historical time series as a
standard
... We'll have to think about how to do that
... We'll do with prototypes in the plugfest first before we put in
a normative spec
... We do need to have geolocation data encoded in a standard
way
... We'll include Christine Perry from OGC
Kaz: It sounds like ITU-T guys always
assume a digital twin kind of service
... They do not think about direct peer to peer?
McCool: That is true
... kaz: we need further discussion about that, though about linked
data
... Physical objects, web service
... Concepts that is included in the information model
... Object oriented factory
... Do they mean RDF triples and stores information model?
... I believe it is an information model about objects
... Smart home use cases
... Similarities to Mozilla architecture like a home gateway and a
dashboard
Lagally: The underlying model of these
specification is manage devices in a user's home
... Central authorization and access control
... Should revisit sometime
McCool: There is a thing that manages
secrets and handles access and authorization
... Idea of access control
Lagally: Let's continue this in the next
call
... So that Sebastian can cover OPC-UA
McCool: Next step is to confirm from ITU-T if I represented their intention correctly
<cperey> <waving>
<inserted> PR 90 - Joint activity for a standardized OPC UA Binding
Sebastian: Sharing screen for Joint
activity for a standardized OPC UA Binding for the web of
things
... Many people are not aware of what OPC UA is
... It's very industrial
... This is a very permanent protocol in the industry
... WoT is open to adopting different protocols
... We have an existing liaison with the OPC UA foundation
<kaz> OPC liaison
Sebastian: We have to write a joint
charter for this topic so we can standardize things
... I think the work effort should not be huge
Lagally: This is important to do, and it's
good we have a formalized use case
... We should come up with a charter of reachable goals
PhilipTran: Question on ITU-T
... will a centralized WOT server be supported in the current work
of W3C WOT?
Lagally: It has been demonstrated in
previous plugfests
... The current architecture supports that and it has been
demonstrated
... We an industrial home scenario
Kaz: I'd like to ask you to generate a summary of your intended draft charter, sebastian
Sebastian: Yes, we need an agreement -- I can make a proposal of a first charter draft and we can use it to make a discussion with W3C management
Lagally: Thank you
Kaz: Usually when W3C like to
generate co-authored standards with another standards organization,
we need a memorandum
... your initial draft can be the basis
Sebastian: Question on membership, if it's required to be OPC UA member to establish this join working activity? The answer is no.
Kaz: What it sounds like what you want is a relationship like what we have between W3C and OGC
<McCool> My issue was about the need to generate a description of the UC task force for the web page; I created an issue: https://github.com/w3c/wot-usecases/issues/92
<inserted> scribenick: kaz
<kaz> PR 89 - add HTML template for use cases
Kaz: can be safely merged :)
Lagally: yes
[adjourned]