W3C

- DRAFT -

WoT f2f - open day

20 Apr 2015

Agenda

See also: IRC log

Attendees

Present
many, many, people
Regrets
Chair
daveR_and_joergHeuer
Scribe
tidoust, fsasaki

Contents


Web of Things Framework, Dave Raggett, W3C

<fsasaki> dave is presenting http://www.w3.org/2015/04/w3c-wot-framework-munich-2015.pdf

<Joerg> yes I will do it and query after Dave's talk

<kaz> scribenick: kaz

claes: there were different options on the slides
... on the possible servers
... we can expect IPv6 as the basis of device identification
... what is your view for the security?

dsr: security problems for IP devices
... we should look at layered security approach

Joerg: due to security approach, what should we consider?
... integration task
... we should be open to which architecture is used for security

dsr: we'll be discussing security tomorrow
... having a task force for the discussion

wonsuk: physical web by Google
... want to know about the difficulty which that
... what do you have in mind?

dsr: conflict with IoT Asia
... Google Physical Device fits well

wunsook: bluetooth work at W3C
... no problem to work for the bluetooth api at w3c?

dsr: there is a CG
... if the web browser is the UI for WoT
... the browser can't solve security issues by itself

presentation by Harting

Frank's talk

(presentation should be added later here @@@)

(4/16) 25b devices

(5/16) what is the key technology?

(6/16) Preservation of status quo and safeguarding the future

(7/16) Elements of IoT in the physical layer

(8/16) Innovation

(9/16) Integrated Industry

Trends and influences and technical potential

(10/16) Individualized World is changing...

(11/16) Individualized World

power to signals

all in one connector

providing not only those products but machines to make them as well

(12/16) Mass Customization needs for individualized solution

modularized manufacturing systems

hardware and software

both are modularized

Vertical integration with business environment

Efficient reconfigurable system: business level and execution level

Customer interface

(13/16) HARTING IT System Integration

Components: RFID Devices, Modular Embedded Platform

Linux-based

Software: Event stream processing middleware

(14/16) Vertical Integration - Cyber Physical System

(15/16) Vertical Integration Szenario

Customer-> Product configuration-> Check availability

security is one of the key topics

modular approach is useful to flexibility

the boxes include multiple processing capabilities

Joerg: various application areas
... how to integrate them?

frank: information on what kind of connector is managed
... all the information can be integrated with the connector itself or the center
... can use the application at the center as the manager
... we have much more control for the devices
... controlling modules
... possible to detect unexpected behavior of the modules

Home automation use cases and requirements, Kazuo Kajimoto, Panasonic

(slides to be added here @@@Panasonic)

(2) Wonder Life-Box 2020

Tokyo Olympic games will be held in 2020

(demo video on "Your Life in 2020")

The mirror detects your health condition.

Smart delivery service using the smart locker.

"House Concierge" manages uses' needs.

also recommends some plan on travels.

can detect when the users want to sleep and turn off the light.

(3) Use Cases

(4) Home Automation Categories (1)

air quality, lighting, audio visual, home energy management

(5) Home Automation Categories (2)

home security, cooking assistance, wearing advice, beauty advice

privacy issue

should enhance owners' authentication

(6) Home Automation Categores (3)

wellness/healthcare, home delivery management, emergency mode, in house moving assistance

CE manufactures can help related business suppliers, etc.

there are many disasters in the world

automatically show the evacuation ways

(7) Home Automation Categories (4)

dialogue/concierge system, safety and cyber security system

a wot device and an iot device might use different technologies

currently a GW doesn't have strong capability for universal conversion

smarter GW could let people handle any IoT devices as if they were WoT devices

combination of technologies is important for security as well

wonsuk: two questions
... 1. is Panasonic preparing some products for IoT
... what kind of protocol are you using?
... 2. there are lot of stuff for the "smart house"
... what is the key?

kajimoto: there are many protocols used currently
... including ECHONET
... we should understand at least 3-4 major ones
... but it's difficult to have only one specific protocol
... because it depends on each device
... so GW should identify at least all the popular protocols
... 2nd question
... we should categorize all the use cases
... we have a very big menu
... a user would use some of them
... currently our system is implemented by only Panasonic
... applications could be categorized into some ways

benedikt: design centrally? or decentraized?

kajimoto: all the information is collected by the central server
... in the future (e.g. 2020) each terminal could become even stronger, and might be going to have capability for distributed system

Joerg: how much cost?

kajimoto: cooperate with house makers
... they sell whole "smart house" as their product
... that is one possible solution
... how to combine with our cloud service is the challenge
... good topic for W3C as well

Use cases and pilot projects for the COMPOSE platform, Charalampos Doukas, CREATE-NET

(slides to be added later @@@COMPOSE)

(2) COMPOSE Platform

platform as a service

service discovery

data provenance

(3) Need for Web of Things

Things (Control, Data Collection) <=> Services

(4) Need for Web of Things (contd.)

main features for IoT

special features

more than GET request

there are already devices connected to the Internet

(5) The Challenges

Things/Control/Data Collection/Semantic Description <=> JSON Schema

(6) The COMPOSE Solution

The Web (Service) Object

(7) Use Case 1 - Smart Spaces

what: samrt retail analytics, BLE Beacons on shopping carts, etc.

where: coop supermarket in Trento, north Italy

(picture of the UI for that system)

(8) Use Case 2 - Smart City

what: optimized car pooling

where: city of Tarragona, Spain

(9) Use Case 3 - Smart Territory

what: ski information app

where: Trentino region, north Italy

cd: actual scenario in Trentino

http://go2ski.eu

johannes: how to create the objects and how to mashup them?

cd: semantic annotation
... find service already available
... how to combine which and which
... aggregate data
... we use Node-RED
... for navigation

short panel

Joerg: there are still 10 more mins
... maybe we could have some discussion on the presented use cases a bit

(Dave, Frank, Charalampos and Kajimoto)

Joerg: your need for actions?

cd: building platform for developers
... as much general as possible
... actuator should be also considered
... security and access control is challenge
... open and flexible scheme is needed

frank: horizontal platform for vertical industries
... @@@ missed Frank's words

kajimoto: cultural issues in Japan
... population is decreasing
... many older people
... need to provide solutions
... also need to think about the technology trend
... we can think of use cases for that purpose

university@: source of security/privacy requirements?

scribe: governmental requirements or customers ones?

cd: depends on the domains
... medical data depends on regulations
... in our case, based on the user access control

frank: very limited
... somebody hands over the data
... conditions depend on the situation
... security is extremely important
... the business model changes completely
... final users own the data as well as the products

Joerg: we discussed a lot

kajimoto: home automation use cases
... one is cloud-based approach
... all the devices are controlled via the cloud service
... another possibility is direct connection
... generic architecture would be better

cd: thinking of edge computing as well

frank: sending data based on the IP address
... data combination
... for existing machines
... additional information could be used

Joerg: tx

brief descriptions on demos

ken: NTT Communications
... two demos
... one is the one on Chrimen project of Mozilla
... will do for Mr. Takagi from KDDI this time
... mozOpenHard project
... open-source computer board

eduardo: Oxford Flood Network
... collecting data
... and analyze it
... "nominet innovation"
... (picture of a sensor)
... send realtime data
... analyze the data realtime
... live river levels
... monitor sensors

soumya: research engineer from Eurecom
... lightweight M2M gateway

saki: NTT Communications
... telepresence robot control via WebRTC
... human like robot
... we can control it (in Japan) from outside Japan
... WebRTC platform named "SkyWay"

sebastian: (no slides)
... Siemens
... micro controller
... using IP-based communication

jonathan: ETRI
... WoT.js
... WoT appliation framework
... three kinds of reason
... generic approach for WoT
... solve technical issues
... find new architecture model
... conseptual model
... WoT.js
... consists of: note.js, express.js, etc.

Joerg: now break!

[ break ]

next session will start at 11:30

<tidoust> scribe: tidoust

Web of Things architecture and use cases (Soumya Kanti Datta, Institut Telecom)

Soumya: Preliminary, IoT was about connecting things and proposing services on top of that.
... We have all seen predictions of volume, 50 billions by 2020.
... What the picture does not show are the challenges: several incompatible deployed platforms and silos
... An electrical engineer of mine often wonders why he would need to use his smartphone to switch on a light, instead of the usual button. We need to understand what the user needs.
... No uniform nomenclature for sensors. The temperature could be written as T, t, Temp, etc.
... This makes it hard for application developers.
... The Web of Things is emerging as a valuable solution to address these challenges.
... Leveraging existing Web standards, exposing functionalities through RESTful APIs to make it easy to GET values.
... I gathered requirements for Web of Things. We need a uniform desription of devices/things. Then discovery, including P2P, which is very ambitious as it's about exploring the social aspect of WoT. Then a very important part is the management of devices.
... Sensors may switch from one network to another for instance.
... Provisioning, management, end user access control.
... Then mapping to existing protocols HTTP, CoAP, etc.
... Security, Privacy, Trust, of course.
... One core topic for the group, I think, is data management and repository (DMR).
... Data processing enable "smart" things to take smart decisions.
... Looking at the WoT Architecture that we're considering [see slide 8]
... At the very bottom, you have devices. Then gateways, networks and mobile clients connected to Web of Things server providing DMR, access rights, discovery, etc.
... One idea would be to move lightweight parts to the gateway since we have more powerful gateways nowadays.
... Proxy-in (sensors) and Proxy-out (actuators) approach to enable the creation of virtual instances of physical devices. Each proxy has an URIs. Proxies can support both smart and legacy devices.
... For description of devices, I re-utilise CoRE Link Specifications, in JSON, with a proposed uniform nomenclature.
... The good thing is that description is very lightweight (less than 1Kb), working on JSON-LD integration.
... Device management is based on OMA Lightweight M2M Technical Specifications. That defines a framework with several benefits: provisioning, registration. Taking the example of a Thing announcing itself every 5 minute. Server knows that the thing is no longer attached if it does not receive an announcement after some time. Configuration update can help change the value.
... Data processing is about making sense out of raw data to derive actionable intelligence. 3-step solution:
... 1. add side information to sensor using SenML. For instance, room temperature? outside temperature? In Celsius, In Fahrenheit?
... This gives the context
... This first step produces metadata
... 3. Derive smarter decisions.

s/... 2. Decorate the metadata with semantic reasoning, based on knowledge of domains, to create a high level abstraction.//

Soumya: M3 approach, implemented recently in ETSI OneM2M architecture.
... For End User service creation, it can trigger action if a pre-defined event occurs (someone enters a room leads to the light being switched on)
... Remote and P2P discovery are work in progress, along with DMR and security.
... Now looking at use cases. The Smart Home use case is the main one [presenting slide 16]
... The cross-domain use case is very interesting to consider [presenting slide 17]. The user could be suggested iced tea if weather is hot and relevant ingredients are available. Very simple idea but very powerful as well.
... Another example: a vehicle could automatically swtich on the fog lamp based on weather sensor.
... In summary, I discussed our motivation and requirements, presented an overview of a WoT architecture. I'm interested to lead a WoT Architecture Task Force

Henrich_Pöhls: What do you mean by end-to-end security. Where are the ends?

Soumya: Between the physical device and the mobile client. It's not only about having HTTPS in-between. From a mobile client, it's important to know that you can trust the provider of the gateway.
... We have to have a way to establish the trust on the device, the gateway, the network.

Dave: @@?

Soumya: By vocabulary, I mean the uniform nomenclature.
... Developers need to follow the guidelines provided by W3C very strictly.

<inserted> scribenick: tidoust2

Soumya: In terms of implementations, there is embed-JENA(?), which takes some processing power. The latency might be 1 or 2 seconds for small devices.

Dave: I was talking with someone from Cisco about similar issues.

Securing the Web of Things (Daniel Schreckling, University of Passau)

Daniel: Context is the COMPOSE EU project, looking at solutions we implemented. We have a bunch of stakeholders that want to communicate.
... Instead of redoing things again and again, we want ro reuse things and combine services. This is more software focused.
... Keep in mind that this can also apply to devices.
... In the IoT, something that did not explicitly show up in previous discussions is that we're sharing sensors across applications.
... You have unpredictable interactions between applications in particular.
... What do you do with security in such a context? Different applications have different security requirements.
... What we're currently deploying is a very static access control mechanism.
... That may not cope well with IoT needs where you want to share.
... The COMPOSE approach is to go down to the data leve and implement security at that level.
... One of the things that you need is identity management. In COMPOSE, it's attribute-based which provides a lot of freedom. For instance, you can assign an OAuth attribute to devices at home.
... This is not currently possible with deployed IoT systems so far.
... Showing an operational view [slide 6]. Two weather sensors provide values. We need to associate security, policies, reputation. Same thing for a service, who is able to assign the data, what kind of sensors you want input from, and what kind of contract.
... When data flows from the sensor to the service, you also need to associate flow policies, security state and provenance.
... Which states did that value go through?
... The idea of flow policies for data is that you use some JSON document tagged to data, defined over actors and you can create classic read/write rules.
... You can view that as a set of locks. As soon as a lock is closed you cannot write the data anymore.
... On top of that, you have contracts that describe what the service is going to do.
... We derive them automatically. They are refinements where a developer might say "I know the constraints that my software has". The goal is to be able to blame the developer later on if the contract is not fully respected.
... Going back to example combining a weather service and a service that provides places to meet, we need to ensure that conflicting constraints are handled properly.
... In COMPOSE, we extended Node to handle our security policies, to ensure that legal flows are allowed and illegal ones are not.
... We can also compose things and analyse the composition itself. We may use TAKS and Klee to analysis the JavaScript and native code. This might reveal some misconfigured things, which would lead to composition reconfiguration.
... if you think back to the original solution where we only had node monitors, this is not so efficient, of course. On other hand, we could also put monitoring directly within the code, and split up the security components, which saves the centralized approach and allow us to compute reputation and manage provenance.
... Reputation and provenance generate a lot of data that can be fed back in the system to adjust policies: "I only want data that was generated by Google sensors, not by others", etc.
... What we don't address is the security of the physical devices themselves.
... Also it does not address tampering if someone introduces a sensor in the wrong place in the production line.
... The manufacture also has to guarantee that the firmware cannot be hacked.
... About data confidentiality and integrity, everything goes to the cloud. You can decide yourself how much goes to another peer.
... Note that the analysis tool and instrumentations are somewhat limited. We can do Java and JavaScript but other languages are not yet supported, and you should not forget that these systems are not perfect.
... They might be wrong. However, it's not as bad as static access control.
... Conclusions: it seems the mechanisms that we introduced based on flow control frameworks appear to be a perfect match for most use cases and create new dimensions for policy enforcement.
... Of course, these mechanisms introduce storage and processing overhead, and something that security experts usually do not like, meaning a dynamic security enforcement architecture.
... Final question is: does the Web of Things want to face this fine granularity? Are the gains worth the cost?

WoF Use cases and solutions at FZI by Benedikt Kämpgen, FZI

Benedikt: Comparison between the Semantic Web and the WoT. The WoT focuses on bits and bytes, real-time and dynamic domains.

[Benedikt going through slides, presenting FZI]

Benedikt: First use case I wanted to touch upon is Ambient assisted living use case. We want to use intelligent sensors for elder people and people with disabilities.
... How much efforts need to be put to enable this use case?
... Going through an example. The stakeholders are the handicapped people, nursing staff, relatives. Things are wheel chair, smart phone, all sorts of sensors. What benefits from linking between things?
... Estimating the action that user wants to execute for instance (pointing at a light to switch it on)
... Benefits from interactions between things? Privacy is improved because a restricted amount of information leaves the house.
... Energy management could also be improved by things being able to communicate their battery status.
... [presenting AAL scenario, slide 9]
... The context management could device what is the perfect medium to remind the user about taking a medicine depending on the context.
... Another use case: Digital Entreprise. The goal is to transition from sensing entreprises to proactive enterprises, e.g. by identifying machines that are likely going to break soon.
... Such system must be scalable, distributed.
... From linking these things, you could understand relationship between the behavior of things and the KPI.
... About possible contributions, we are using semantic mediaWiki to keep humans in the loop. The openAAL tool that I presented.
... Linked APIs are a way to describing services.
... We have also been working on rule-based languages such as Linked-Data-Fu.
... To answer W3C questions about interest in WoT: I think it's mainly about connecting things together based on mechanisms that have evolved over the last 10-15 years.
... Stream data protocols, rule enginges, service description languages would be technologies of interest.

[ lunch break ]

<Alan> scribenick: Alan

Social this evening

Jeorg talks about networking

Jeorg: Idea is to meet at the underground station at 6P
... We'll get together then and then take the subway
... [explains how to get there on your own]

IETF IoT related standards, and the Thing to Thing Research Group,

Carsten: I'm going to talk about standardization on the network side
... it's been 10 years of work

<inserted> scribenick: JAB

Carsten: Let's start by defining IOT - was originally RFID
... We're expecting 50 billion nodes by 2020
... It's a matter of scale that we know how to do. We have very little to change in this space.

<fsasaki> scribe: fsasaki

Carsten: constraints of nodes lead to constrains of the networks
... nodes work on batteries, need to have them sleep a lot
... there are various networking technologies
... we can use a stable internet protocol
... constrained node networks
... terms for this: IoT, wireless embedded internet, IP smart objects, ...
... wireless sensor networks: rather about research - designed for a specific research grant
... "designed for grant proposals versus designed for decades" - the latter is the goal of the ietf work
... many people in IETF try to build architectures that it works without the cloud
... can we move the baggage? you can make it work but may not want to
... two camps: "Ip is too expensive" ... vs "ip works as it is - why change?"
... both are right: boundaries between IP or not IP are being moved
... 7 WGs in the IETF are working on IoT
... group 8 is created soon
... several groups related to security, one related to applications
... layer of IP on constrained devices by a given company (ARM)
... comparison of hype IoT and real IoT: different technologies being applied
... 2008 starting working on routing
... now on applications: group on CoRE - constrained restful environments
... CoRE done in basic mode. Elements of the web: HTML, URIs, HTTP
... HTML in IoT still needs work
... URIs for IoT is suitable
... HTTP still needs work, e.g. for compression related to IoT
... CoAP can use rest proxies
... run proxy chain from CoAP to HTTP
... CoRE embraces REST. Adds "observe" and web linking
... details see http://coap.technology
... security in CoAP is not optional: DTLS 1.2
... since May 2015: ACE working group
... on authentification and authorization
... since 2013: CBOR. Representing JSON in binary
... http://cbor.io/
... next step: COSE (IETF JSON security)
... need to work together, also with W3C
... cooperation items: application layer technologies like mgmt for constrained nodes, security, ...
... now switching to IRTF (research arm of IETF)
... first wave of IoT standards done
... IoT consortia are filling gaps

<scribe> ... new requirements for research coming up

Claes: how are unique ipv6 addresses allocated?

carsten: has a privacy element
... unique identifiers can be used for linking, on IP and net layer
... recent work around with randomized MAC addresses
... one question - how persistent do the addresses need to be?
... we de-emphasize DNS
... there is no problem with having the same address in multiple times

Benedikt: you mentioned the "observe" command - can you explain why?

carsten: sometimes a server has new information and wants to let the client know about that
... a small addition to the architecture, integrates smoothly into REST

richard: standards above coap
... what is your opininion on having standards on the resource description layer?

carsten: personal opinion: there are a lot of pits here, e.g. doing the next wsdl
... with the web linking based architecture we are not done yet
... e.g. there may be other metadata that one may to add
... interesting thing of web linking is that easily you can add new information, just document them

Semantics for discovery and interoperability of services, Carlos Pedrinaci, The Open University

<JAB> sccribenick Alan

<JAB> sccribenick: Alan

<JAB> scribenick: Alan

Semantics for discovery and interoperability of services

<inserted> scribenick: JAB

Carlos: I'm going to be taoking about aspects of the last question
... This is a mapping of some of the solutoisn and standards that are around
... Focus on how to make this work
... Verticals slide
... It's harder when you look across domains
... What ususally happens is they go for an API

[Horizontals slide]

Carlos: The stand is on the Web part of WoT
... The win is to find the way to repurpose components

[Find Understand Intereact Combine slide]

Carlos: "this needs to happen despite Vendor Specific silos"

<inserted> scribenick: Alan

Carlos: behind the scenes there will be different things, but at the application level you should be able to work together

[xively slide]

[dweet.io slide]

[ThingSpeak slide]

Carlos: I was trying to figure out a way to show this.
... It brought me back to talks about mashups from 2007
... That has worked, and in my opinion the WoT needs to head in the same direction
... Looking bat at Web APIs slide
... we should rely on developers implementing things properly but that's a problem
... but that's not the case, even HTML pages aren't implemented properly
... Semantics are essential, you need to consistently say what the data is, etc.
...
... The SAWSDL standard only focuses on three things


.[picture of develper environment]

[Description slide]

Carlos: We've been working on these things and have solutions.
... It's university research but it works
... we have iServe covering this
... we use existing datasets with the definitions that are out there.

[data mining slide]

[Describing Things slide]

[Things Functionality slide]

[Things Interface slide]

[Discover slide]

[iServe demo]

[Interact slide]

[screen shot of data extract]

[Combine slide]

[Composition slide]

[ending slide with links]

IoT platform glue.things - Node-RED as basis for authoring tools

Robert.Kleinfield

Robert: We try to provide solutions to mashup IoT / IoS
... Our proposal is we have a platform galled glue.things
... We need to identify the building blocks for WoT

[What is Glue.things slide]

[glue.things overview slide]

Robert: It's important to focus on the data depicted on the right side
... There are various solutions for data management, we try to provide integration for them
... We've integrated the red, I'll explain how we did that.

[how can applications be developed slide]

scribe: Application in 3 steps - 1. Connect, 2. Build and 3. Distribute

[glue.things dashboard slide]

[glue.things dashboard detail slide]

Robert: Each device is defined in data management and the streams are represented in the mashup

[glue.things composer slide]

[example of My Composer screen]

[Marketplace slide]

[My Automations screen shot]

[Developer Tools slide]

[Technology Stack slide]

[W3C Web of Things slide]

[ETSI M2M slide]

[Conclusion slide]

[Thank You slide with pointers for info]

<fsasaki> scribe: fsasaki

discussion

dave showing thing description from his presentation

http://www.w3.org/2015/04/w3c-wot-framework-munich-2015.pdf#page=14

dave going through json-ld description containing events, properties, actions (not always given for each thing)

dave: now things as an agent

http://www.w3.org/2015/04/w3c-wot-framework-munich-2015.pdf#page=15

dave: behaviour here expressed via JS - but that is up to the server
... here event is send and bind to a function
... calls unlock action to the door
... many of these things have to be thought throuh

see http://www.w3.org/2015/04/w3c-wot-framework-munich-2015.pdf#page=16

dave: see for efficient transfer of structured data
... ways of specifying vocabularies
... binding protocols may be a standardization item
... standardizing servers - normally done via protocolls, e.g. http / coap, so not so much to be done
... provisioning and life cycle mgmt
... discovery
... discussion about security and privacy
... legal terms - can we borrow s.t. from the web like creative commons?
... some thoughts - now everything open for discussion
... question to stefan on xmpp - how hard would it be to define restful bindings?

stefan: there is already s.t. that helps with that
... I would suggest to come up with new items that specifically target WoT
... xmpp is extensional protocoll, core protocoll only defines stream of xml that is being exchanged
... we would have to come up with new items to extend xmpp

felix: asking about modeling domains more general or specifically - how to interrelate that, e.g. for end customer and manufactuer?

dave: example of different manufactures who have different approaches
... project involving etsi - worked on interrelateing vocabularies
... w3c is not in position to define vocabularies, but can build foundations to build relations between vocabularies

Carlos: experience from SAWSDL: we defined a hook how to "attach" semantics
... similar to REST principles
... establishing / building principles may help
... could be some space for defining a core ontology
... but then there are more specific domain specific items
... see the example of schema.org which is the "upper model" in that way

dave: important to take the different models of abstractions into account

alan: this is open discussion session
... we need people from the various industry involved e.g. to move a task force forward
... to move s.t. forward you need two implementations to create a standard
... see e.g. fraunhofer showed something which can be input to this - if you don't want this to become a standard, join the conversation and make your voice

dave: thanks - there is a deliverable for the group that will help with gathering the use cases

jörg: this morning was interesting to see: different proposals on architectures

scribe: e.g. architecture of a hub see glue.presentation - then e.g. see architecture of network of things
... coming to agreement on the set of architecture and then discuss building blocks seems to be pre-condition for the IG
... then the question is: what other pre-conditions do we have for the IG
... then some questions to think about: what are the most two important building blocks?
... security was the most refered topic today
... but not sure if security is a building block
... whatever we discuss - need to understand if it will be a building block
... also "taking care" can mean: we use something that already exists
... see e.g. CoAP / IETF relationship
... so in IG we need to discuss: pre-conditions, building blocks, what do we want to take care of

Henrich: you may need secure building blocks
... be precise in terms of the endpoints that you want to communicate with

<kaz> s/Daniel_Shreckling:/Henrich:/

dave: need a group of people - what are the right questions to ask
... there will be some parts done, e.g. some metadata items that already can be described

jörg: technical aspect important, but also: how to work on it in the interest group?

gisela: it depends on the scenario that you want to build
... e.g. for security & privacy, you need confenditality
... for security there are non-functional and functional requirements
... then also important: standardisation point of view
... there are many standardisation activities that deal with WoT / IoT
... you need a standardisation landscape to doc on other activities
... not only IETF, but also ETSI, ISO for smart cards and there is a new ISO group for IoT

dave: one proposed task force is: to work on liaisons
... to make sure that messaging is clear - but we need some help of people working with that

benedikt: one important task: decide what would be the outcomes of the IG
... e.g. would you want to establish best practices?

dave: part of the answer is: understand what the WoT means
... we don't want to create yet anther platform, but would allow the platforms to talk to each other via the web

<Alan> Felix: Dave mentioned there are specific task forces, do you have them to share?

jörg: won't define what is "wot" but rather go from the buiding blocks

<kaz> list of proposed TFs so far

dave: tasks force being discussed: see the agenda link, security & privacy, some specific sectors
... e.g. question to daimler: is it time to think about automotive task force?

Andreas: security is clearly important
... not clear to me yet: how can the integration into WoT be done?
... that is an important question for the car
... in a town you have connectivty, you get music from the cloud - but how about rual areas?

dave: so e.g. you need intelligent services that can predict than you are driving, using some gsp data, then downloading something

sebastian(bmw): many technologies discussed today relevant for automotive

scribe: automotive business is slow, so we are not ready to join jet
... but interesting discussions. protection of data is important
... then reliability, e.g. funcitonal safety
... in cars low engergy state is also important, you might use in certain situations
... so there is place for tasks forces, but we need more time at bmwi to discuss internally

<inserted> scribenick: kaz

kaz: not sure if we really need an Automotive TF within the Web of Things IG at the moment
... would suggest the WoT IG should collaboratively work with the Automotive WG to see what should be done within the Automotive WG and what should be done within the WoT IG

<tidoust> scribe: tidoust

wrap up

Joerg: I think we can do this in 3 parts
... 1. Look back at what we discussed today
... 2. Look at upcoming two IG F2F days
... 3. Gather inputs from today's participants
... Most comments are from the Smart Home arena, but it was nice to see the integration. Amazing video from Panasonic. We're really wondering how to enable this.
... Use cases around the public space, parking (smart city related), etc.
... Of course, our task is to collect these use cases, but also to start to use them
... in particular to derive requirements for the WoT framework.
... Question is what is common from all these domains?

Dave: W3C work needs to meet the needs of actors. What kinds of use cases will help drive the work? I'd like to ask Alan from W3C to say a few words on that.

Alan: I think that when we look at moving this work ahead, there are a couple of dimensions to look at. One is by the industry, but also use communities.
... I agree with the notion that security in building blocks is needed.
... If there's a topic that you care about, you need to step up. It could be around smart home, or around something else entirely. What does the WoT mean to the mobile space? That could be a task force.
... One group that is not here today, Boeing, has use cases around network at airports and within airplanes, for instance.
... The task forces control their own pace. How fast can we move? That's how fast task forces move.
... It's based on interest that things moves on.
... Someone mentioned 70 IoT organisations that they contribute to, that's one figure too many.
... Managing liaisons is important.
... We work in a Royalty Free pace but we need IPR to build the standards. Your companies have the IPR. Which of those are you willing to contribute so that we can spin off working groups?
... The IG will produce use cases and requirements, no IPR contribution there, but you need to think ahead about IPR for possible follow-up working groups.
... I would encourage you all to raise your hand if there's some group that you would want to start.
... If something is missing, just let us know as well.

Dave: We all heard about how hard the WoT is going to grow in the coming years. We cannot do everything, we have to focus on a few building blocks that can progress rapidly.
... We talk about some of those today, and hopefully we can make them more crisp over the next couple of days.
... Are there things ready to move?

Joerg: Looking at the agenda for next two days, we have to do some work to get us organized and started, that's the use case discussion in my view.
... Also the beginning of the second day fits in that category.
... Besides that, we have the core of our activity, the building blocks of our framework and our landscape.
... I think it will be interested to start with a plenary discussion and then have some break-outs.
... The discussion of which Task Forces and how to organize them will be on the second day.
... Probably we will have time on Wednesday morning to discuss security and other related issues.

Dave: With respect to the number of F2F the IG could have, I heard some companies willing to move fast, which translates into more meetings.
... We're planning one for Summer. W3C TPAC will take place in Sapporo, Japan, end of October.
... How many F2F meetings? What is the best way to work? We try to alternate calls at friendly time for different parts of the world. Is that good enough?

Joerg: To provide some perspective beyond next two days, a question came up about defining the Web of Things. Maybe it's more expressive to start listing building blocks that would make a first version of the Web of Things framework.
... [showing lists of proposed task forces]
... If you want to get involved in discussions of some of these building blocks or want to assign priorities, please say so.

<kaz> list of proposed TFs

Joerg: Even if you do not participate in the next two days, you can review the minutes of these discussions and look at the IG material to understand the focus.

[ End of the open day ]

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.140 (CVS log)
$Date: 2015/04/20 14:56:37 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.140  of Date: 2014-11-06 18:16:30  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/IT/IP/
Succeeded: s/wunsook@/wonsuk/g
Succeeded: s/Fran's/Frank's/
Succeeded: s/them/them as well/
Succeeded: s/RFID/Components: RFID/
Succeeded: s/video/video on "Your Life in 2020"/
Succeeded: s/(1)/(2)/
Succeeded: s/[ break and demos ]//
Succeeded: s/engineer/engineer from Eurecom/
Succeeded: s/micro controller/Siemens/
Succeeded: s/in/at/
Succeeded: s/;/,/
Succeeded: s/kenpen@:/benedikt:/
Succeeded: s/... 2. Decorate the metadata with semantic reasoning, based on knowledge of domains, to create a high level abstraction.//
FAILED: s/... 2. Decorate the metadata with semantic reasoning, based on knowledge of domains, to create a high level abstraction.//
Succeeded: s/He/He_(University_of_Passau)/
Succeeded: s/He/Daniel_Schreckling/
Succeeded: i/Soumya: In terms of/scribenick: tidoust2
Succeeded: s/FZI/Benedikt Kämpgen, FZI/
Succeeded: s/Speaker:/Benedikt:/
Succeeded: s/WoF Use cases and solutions at Benedikt Kämpgen, FZI/WoF Use cases and solutions at FZI by Benedikt Kämpgen, FZI/
Succeeded: s/Daniel_Schreckling_(University_of_Passau):/Henrich_Pöhls:/
Succeeded: s/@@@:/Claes:/
Succeeded: s/YYY:/Benedikt:/
Succeeded: i/I'm oging to/scribenick: JAB
Succeeded: i/behind the scenes/scribenick: Alan
Succeeded: s/...behind the scenes/Carlos: behind the scenes/
Succeeded: s/I'm oging/I'm going/
Succeeded: s/swasdl/SAWSDL/
Succeeded: s/QQQ/Daniel_Schreckling/
FAILED: s/Daniel_Shreckling:/Henrich:/
Succeeded: s/Daniel_Schreckling:/Henrich:/
Succeeded: i/not sure if/scribenick: kaz
Succeeded: s/Jorg:/Joerg:/
Succeeded: s/stefan//
Succeeded: i/Let's start by/scribenick: JAB
Succeeded: s/... Even/Joerg: Even/
Found ScribeNick: kaz
Found Scribe: tidoust
Inferring ScribeNick: tidoust
Found ScribeNick: tidoust2
Found ScribeNick: Alan
Found ScribeNick: JAB
Found Scribe: fsasaki
Inferring ScribeNick: fsasaki
Found ScribeNick: Alan
WARNING: No scribe lines found matching ScribeNick pattern: <Alan> ...
Found ScribeNick: JAB
Found ScribeNick: Alan
Found Scribe: fsasaki
Inferring ScribeNick: fsasaki
Found ScribeNick: kaz
Found Scribe: tidoust
Inferring ScribeNick: tidoust
Scribes: tidoust, fsasaki
ScribeNicks: kaz, tidoust, tidoust2, Alan, JAB, fsasaki
Present: many many people
Agenda: https://www.w3.org/WoT/IG/wiki/F2F_meeting:_20-22_April_2015_in_Munich#Open_Day_.28Monday.2C_April_20.29
Got date from IRC log name: 20 Apr 2015
Guessing minutes URL: http://www.w3.org/2015/04/20-wot-minutes.html
People with action items: 

[End of scribe.perl diagnostic output]