Workshop on the Web of Things - day 1

25 Jun 2014


See also: IRC log


many, many, people
Jörg Heuer, Dave Raggett
fsasaki, Marie, ph, Philipp


<fsasaki> scribe: fsasaki


jörg heuer introduces the workshop

jörg: why is Siemens interested in the WS topic?

scribe: Siemens organized in 4 sectors: industry, energy, health care, infrastructure & cities
... we realized: we need to think across the verticals
... web of thing activity is a lot about crossing domains
... we need to "to understand the elephant called web of thing"
... we'd like to exchange our perspectives and uses of the term

Brief introduction to W3C and workshop aims

dave: w3c: creating open standards under royalty free patent policy
... standardization + pre-standardization activities in business groups and community groups
... internet of things is focusing on sensors etc.
... we hope to form a group opinion what the "web of things" is
... web has been around for 25 years now
... it is the most interoperable platform across devices today
... the web has transformed the way we communicate - web of things will do the transformation hopefully again
... why are we here today? there is fragmentation in the technology area around "wot"
... we want to discuss: what kind of open standards are needed?
... what application domains + use cases are relevant?
... what role should W3C take?
... we had many submissions, thanks a lot for that
... we have panel sessions + break out sessions. Each session leader is responsible for taking notes and for reporting back
... we need those minutes for the WS minutes
... over 100 participants and many major companies - also lots of breakout topics, see the wiki linked from the agenda

session "Sharing experiences, use cases and requirements"

first speaker johannes hund

Johannes Hund, Siemens -- web technologies in smart grid and EV applications

johannes: there is a lot of processes across industry domains, involving multiple stakeholders
... there are key enablers. increasing processing power and communication facilities
... example: smart gird application "virtual power plants" (VPP)
... loosely coupled control over "distributed energy resources"
... main application is the aggregation of data and orchestration
... many stakeholders are participating in the scenario: grid opeator, power plant, metering operator etc.
... operator wants to control the energy flow, but for energy traiding that may lead to contradictions
... we need an automation system
... currently we have only a collaboration system
... we are using chatt protocols like XMPP
... example of vertual power plant with distritributed storage
... antother use case: charging of electric vehicle
... there is the vehicle, charging spot, user, energy provider etc.
... so again many user interact
... for automation a system in the eNterop project was developed
... so what drives wot and what does it enable?
... many stakeholder come from automation point of view and from collaboration
... and some aspects to realize that is possible with existing web technology
... we need to provide an extensible framework
... does not need to cover all use cases
... need to encompass security etc.
... we have breakout sessions that show such activities

jörg: thanks, johannes - questions?

pablo: when we did web services in the 90s
... it was focusing on workflows
... then more interaction came up
... but we stil don't have a good way to push information
... as you said: internet of things is closer to social networks
... we need to realize that the traditional web services don't fit for this

jörg: exactly

scribe: with xmp you can break out the things that are modeled with soap services
... into patterns that fit in applications

Yusuke Doi, Toshiba -- Long Life Web of Things with EXI

Yusuke: my point of view here is as a member of W3C EXI working group
... we expect that billions of devices will be connected to the internet
... one aspect of the volumen: E-Waste
... some part of the devices will produce waste
... there will be some early innnovators, some may last long, some may fail
... we cannot update or re-publish the "big things"
... there are candidate wot standards [list of various standards from ISO and others] - they are defiend in XML / UML data models relying on XML Schema
... these things are not so smart and cool, but we can use them to build reliable applications
... focus on the bottom layer: what are the things, how they can interact, how to assure their sustainability
... some of you may think: XML is not ready to be used on devices
... example of wot application using xmpp (=XML protocol)
... the XML is encoded as EXI, that is: binary XML for interchange
... EXI can be used as alternative to XML. because it is binary the encoder does not need to think about encoding, spacing / newlines etc.
... because EXI uses XML Schema as a source of grammer, it is a state machine that can be implemented in a tiny way
... XML is ready for "things"
... need to define: what a device is, data model of things, communication model
... recent activities from EXI working group
... have published as a proposed recommendation: EXI profile for efficient use of memory
... also grammar exchange format (under discussion)
... example of trick for implementing xml schema based communication on tiny devices
... from a certain point of view XML is just a text
... the interpretation comes with the data model
... certain devices don't need to know all parts of the model
... if we can do that we can have a significant reduction of XML schema or EXI grammar size
... such tricks can help to enable XML & standard based communication on tiny devices
... so: wot may require some standards that we can rely on to build long lasting cool applications
... such a standard will require data model description language like XML Schema
... json is good to but it has less ability to describe data models
... XML is a good approach - it has been created by W3C (= us). This WS is a good place to discuss what we should have
... EXI is one example, it is just a tool to realize wot. Main point is: we cannot make standards for everything
... the actual domain specific standards should be discussed e.g. in OASIS (e.g. for health care = HL7)
... maybe the WS can help to make a meta standard, that should be interoperable


yusuke: depending on the domain, sustainability is more or less important
... e.g. for smart phones less important, for train systems more

Masaki Umejima, Japanese Smart Community Alliance -- ECHONET Lite enables Demand Side Energy Management - IP based and IEC's open standardized interface for home appliances

Masaki: my story is based on nation wide implementation, not research
... I am in charge of smart meters in japan, from technical specification to implementation etc.
... this september we will start adoption in Tokyo area
... in 8 years all japan meters will "speak" IP
... my role is to address parameter adoption
... the japanese parameter specification sometimes receives complaints from international partners
... one must understand that Japan's ECHNOET lite power meter is not domestic
... a very unique approach which can provide many opportunities for you as well
... you can download technical spec of ECHNOET LITE from the web
... ECHONET LITE will be IEC standard. 200 companies are in ECHONET consortium, including oversea companies like IBM
... we have close relationsship also in Asia
... step by step we will cover many areas, so now we must discuss with America + Europe
... ECHONET LITE is a liaison partner in W3C, want to discuss with you how to move forward

<mcf> Agenda: http://www.w3.org/2014/02/wot/agenda.html#agenda

[Demo about ECHONET Lite]

masaki: ECHONET commands to turn on / turn off light

. now batch control of air conditioner

s/\. now/.. now

[air conditioner is talking]

masao: about HEMS interop test centre

<scribe> .. ongoing interop test case #1

<scribe> .. new business model "smart house"

UNKNOWN_SPEAKER: being commercialized
... we need to discuss open standards like ECHNONET Lite on the open web platform

philippH: why is it LITE? What was the HEAVY counterpart?

masaki: ECHONET did not support IP
... and another difference: ECHONET was not open, E. LITE is

Nick Allott, impleo -- IOT interoperability using web technologies: lessons learned and future challenges

<mcf> Nick Allott

<mcf> (nquiring minds)

nick: three contexts: webinos project, niquiring minds, and ubiapps
... IP licensing costs in radio technology - currently too expensive
... you as IOT operator don't have IP support since it is too expensive from power and licensing perspective
... 2nd comment: speed is essential
... 3rd point: if you talk about IoT, interop - you should re-use what is available
... we need to see what is good, put it together and address the challenges
... IOT challenges: first connecting devices
... using TLS gives you security, encryption etc.
... second: provenance
... we are using PKI. every connection is based on a mutually authenticated connection via TLS and using PKI
... third: access
... need to be able to control data access
... we picked XACML from OASIS
... then: addressing (remote)
... how to address the sensor / capability on a different device
... the URI needs to be clever - not every domain will resolve to a specific IP address
... the sensor only sometimes may be connected
... so need to be smart on how to produce an URI for this situation
... fith: discovery
... a lot of stuff is available
... since IOT is constrained we need to be careful. we are using FEATURE-URI
... nice thing about it: it is extensible, URI based - everyone can specify their needs
... then: capability
... I as an APP developer - how do I act on data
... the developer wants a Javascript API
... you can talk about IOT protocols, there is a lot of tech out.
... the only way to manage this diversity is to have a functional abstraction
... then: invoking: nice to call a capability from a browser
... for IOT to work: I want to invoke from other systems
... we are using JSON-RPC
... it allows to access a sensor between devices
... with total flexibility of using javascript
... advanced topic: address local, global protection
... policy sync
... handle legacy issues
... at the end there is no single point of control in the architecture
... this is one of the biggest task that W3C could deal with in the IOT space
... if W3C will engage in the space, how much of the problem will be in scope? that is what I want to know at the end of the workshop


jeff: what are the 2-3 things we should focus on quickly?

nick: simple javascript abstraction
... and a security model
... I have seen things fail if people don't address that upfront

demo elevator pitch

[short summaries of demo slots]

now coffee break + demos

<kaz> scribenick: kaz


From things to the Web of Things

Shigeru Owada, Sony -- Kadecot: Android Web API Server for Home Appliances and Sensors

-> http://www.w3.org/2014/02/wot/papers/owada.pdf Paper

-> Slides tbd

shigeru: (Polling Manager)
... polling manager publishes how many apps are interested in each resource
... (WebAPI)
... WAMP protocol
... an application named "Moe-kaden"
... IoT service and cute character
... (Demos)
... two videos
... 1st, RPC demo
... recognizes your voice input
... 2nd, PubSub demo
... distributing manuals
... ECHONET Lite emmulater
... in case of errors, manuals will be shown interactively


ken: kensaku komatsu from NTT Communications, Japan
... home server
... discover the home network server?
... what kind of mechanism?

shigeru: device will be automatically discovered using ECHONET Lite, etc.

ken: but you use WAMP as well?

shigeru: one possibility is using UPnP, etc.

ken: so the Web browsers should use UPnP, etc.

shigeru: right

<mcf> Ryuichi Matsukura, Fujitsu -- Service platform with Web based interface to control devices

Ryuichi Matsukura, Fujitsu -- Service platform with Web based interface to control devices

-> http://www.w3.org/2014/02/wot/papers/matsukura.pdf Paper

-> Slides tbd

ryuichi: (Background)
... devices connected to network
... home security, healthcare, HEMS
... a "service platform" manages the interaction
... (What is the service platform?)
... advantages
... without the service platform, apps depend on different device interfaces
... the service platform makes it easier to use multiple devices
... (Deployment of service platform)
... two types of deployment
... aggregate type and distribute type
... (Functional architecture)
... three categories
... ITU-T SG 13 work
... will become a standard by the end of 2014
... device adapters
... three categories: app execution, device operation, remote management
... if there is a trouble, notice will be sent to the user/admin
... (Devices applicable to service platform)
... basic devices
... air conditioner class has properties like operating status, operating mode
... basic devices support std interfaces like ECHONET Lite, KNX, SEP 2.0
... (Basic device operation)
... home gateway converts device commands
... virtual devices correspond to objects
... s/objects/device objects/
... (Device operation without device object)
... non-basic device can also get connected
... (Sample applications for service platform)
... 24 facilities with 28 kinds of 820 devices
... ECHONET Lite is used
... (Conclusion)
... service platform can flexibly connect with multiple services and multiple devices
... WoT framework should support existing devices
... (Video demo)
... Web interface
... push "living room"
... realtime video image
... remote interface for the air conditioner
... would show the details later


dsr: discovery?

ryu: ECHONET Lite has that mechanism as well

(dsr checks where the memory stick is)

Ricardo Morin et al, Intel -- Programming Device Ensembles in the Web of Things

-> http://www.w3.org/2014/02/wot/papers/ricardo.pdf Paper

-> Slides tbd

ricardo: programming device ensembles
... (Today's IOT programming model...)
... embedded computing
... device centric
... statically partitioned
... constrained by device capabilities
... vendor-specific solution
... scale limited
... (WoT model should evolve...)
... ensemble programming
... everything is a resource
... sensors, actuators
... devices and capabilities
... app runtimes and app services
... apps are dynamic
... resource discovery
... multiparty authorization
... remote workers
... e.g., owner of the building, sensors, etc.
... off-loading computation

those are three particular areas to consider

scribe: (Discover "nearby" sensors)
... need APIs
... hide the details
... intelligent container
... location sensor for cargo
... URL for resources
... could be global advertisement
... discovery phase must be semantic
... concept of "near"
... meets the criteria
... the semantics of "near" includes "local"
... what is the permission?
... user managed access
... fundamental hypothesis is "you don't know anything"
... (Off-load intensive computation)
... we propose using workers model
... (Putting it together...)
... intelligent container
... (Our tam is ready to participate and contribute in these areas)
... discovery, authorization , code off-loading


jeff: tx for interesting presentation
... well-defined areas

ricardo: UMA is IETF standard
... publicly available

jeff: remote support?

ricardo: yes

dsr: demo during the breakout

Dave Conway-Jones, IBM Conway-Jones, IBM -- Node-RED: a wiring tool for the Internet of Things

-> http://nodered.org/ web site

<jeff> Jeff: Are some of the new areas such as remote support available in a public document

dave: (Node-RED)

<jeff> Ricardo: We have worked on such areas, mostly it is error handling (e.g. no resource available) that needs to be added to the Workers spec. We can provide that.

dave: (Why Node-RED)
... IoT does not have a one-size-fits-all solution
... (We need tools...)
... Node-RED can fill that up
... visiting the web site: http://nodered.org/
... (actually localhost)
... (draw a diagram and generate JavaScript automatically)
... e.g., sending an email
... (Node-RED is)
... app generation tool
... (Node-RED is NOT)
... an enterprise strength application runtime
... (Inject node, Debug node, ...)
... (Node-RED animation)
... (From the edge of...)
... (Function Node)
... (Easy to wrap any npm module into a pallete node)
... (demo on localhost again)
... IRC node to WebSocket
... works and easy
... (Live Wordcloud)
... (Live Dashboards)
... (From the edge of the network...)
... (... to the cloud - IBM Bluemix)
... easy to move flows
... (Wiring the internet of things

Andreas Harth, KIT: need parameter description?

scribe: how to maintain?
... who defines?

dave: actual messages are described here (type and ID description)

Andreas Harth, KIT: have to convince people to write the descriptions?

Johannes Hund, Siemens: schema for input/output?

dsr: can talk during breakout sessions

Jon Nordby, The Grid: working on protocols

scribe: interested to talk with you

Kazuo Kajimoto, Panasonic -- Consumer Electronics and Web of Things

-> http://www.w3.org/2014/02/wot/slides/kajimoto.pdf Slides

kazuo: (WoT Applications around CEs)
... many possibilities
... smart house example
... (Wonder Live Box 2020)
... smart home business globally
... working with house manufacturers
... using ECHONET
... the system works with the other companies' devices as well
... convenient, relaxed and high quality life in the near future (in 2020)
... (WoT Model Description and APIs)
... abstract architecture
... right side is the physical world
... vertualization of devices
... application cloud APIs to control vertualized devices
... that is type 1 interface
... type 2 interface is device APIs
... there are those two different types of APIs
... (Comparison of App-Cloud API and Device API)
... cloud might be an agent of device
... both types are important
... we'd like to contribute to type 1
... (Physical Control Protocol Independency)
... many physical control protocols
... but they should be independent from app cloud APIs
... (Variety of App-Cloud API Description)
... our current system is based on JS assuming WebSocket
... trying to support various mechanisms
... (Type of App-Cloud API)
... four types of APIs
... (Concern)
... WoT is our dream
... but there is some concern
... safety issues, privacy issues and business/security issues
... if somebody controls my air conditioner, that's not good
... also an app might be able to peep the others' houses
... user authentication might be better
... better to provide both open API and authority managed API


Laurent-Walter Goix, et al, Telecom Italia -- Smart social spaces: opportunities and challenges of the social web of things

-> http://www.w3.org/2014/02/wot/papers/goix.pdf Paper

-> Slides tbd

laurent: concept is mentioned by Ericsson some years ago
... 2009
... (Smart Social Spaces)
... issues we're facing
... smart objects
... smart social spaces
... (Smart Social Spaces: our vision from IoT to (S)WoT)
... integration of IoT with the Web is limited
... (S.W.O.T on SWoT: so what?)
... activity streams
... each of things is an entity
... identity and relationship
... missing data model
... having discussion on social web of things
... coming up these days
... (Social Web of Things: (some) features & issues)
... cloud-based notifications
... local space notifications
... common issues: SN account vs object addressing, access control ad audience targeting, information filtering, interaction paradigms
... temperature information is not of users' interest
... internationalization
... concurrent (multiple) access
... sharing one thing with many people
... (BUTLER Smart Office trial)
... FP7 BUTLER project
... (Current standardization activities: Open Mobile Alliance)
... white paper on "social web of things"
... (Our interest within W3C)
... web of things wg focusing on "Things"
... social web wg focusing on "social"


Dave Conway-Jones: interesting

scribe: the way works today is each social services are separate
... should work for multiple parties
... that's the purpose of OMA work

[ lunch ]

<marie> scribe: Marie

<scribe> scribenick: marie

[presenting breakout discussions]

<ph> scribe: ph


<marie> scribenick: ph

Object Security in Web of Things - Ericsson

Q (Joerg, Siemens): with constrained devices you may need hardware support - how does this work out?

A: some constrained devices cannot handle all the standards, others can - if device cannot handle, need gateway to handle standards, and prop protocol, which may not provide security

security may not always be needed in local situation

Authentication for the Web of Things (Siemens)

Q (Jeff, W3C): on automotive as best practice - they are hybrid apps - we have automotive bg in w3c - have you looked at that? same conclusions?

A: right now apps are testing the market

in automotive - car industry spending lots of money not for current use cases, but prepare for more

any function car has should be available via api to developers

massive hypothesis, they are testing it with "toy solutions"

Q (Andreas Harth): when i unlock car, why do i have to login to the cloud? why can't i talk to car directly?

A: business decision by car manufacturers - they are sensitive about crm store - not in the cloud, but in the data centers of vendors

Trust-based Discovery for Web of Things Markets (Innova)

<dsr> [note that the monohm talk was cancelled as they weren't able to get here]

Q (Joerg, Siemens): did you look at standardisation?

A: W3C is part of this project - wrt trust: our vocabularies are based on W3C results

also follow linked data approach - from linked data to linked services

Q (Joerg): do you have deployed applications?

A: not yet - testing with three different use cases - two smart cities - one for example on ski slopes - weather forecast from the slopes - also retail use case

Breakout organization, elevator pitches for breakouts

<marie> See the breakout wiki: https://www.w3.org/wiki/Web_of_Things_Workshop_Breakout_Sessions

Andreas Harth (KIT): Scripting in the WoT - how to specifiy interplay between sensors using rules

if sensor x has value y, do z

Dave Raggett, W3C: Service Description for the Web of Things - need way to describe services, what they depend on, what they export ... parallels to linux packaging

Erik Wilde (Siemens): very close to what Dave has proposed - if you had service descriptions, how do you interact with these services - Atom inspired

interaction models for the web of things

also relation to social web wg - and activity streams - can wot infulence what happens in that wg, use it etc.

Alessio (Innova): Business model for WoT - buy and sell objects - some platforms already exist, but are they sustainable?

Jens Haupert (DFKI): Digital object memories for WoT - store information with object throughout its lifecycle

Dominique Guinard (Evrtyhng): The Web of Things Friendly Label - guidelines for device builders and cloud services builders

inspiration from JS world where there are frameworks for building scalable apps

Hauke Perterse (Free University): Application layer protocols and data encoding for constrained devices - want to have end2end connenctivity to constrained devices - look at emerging protocols that are more efficient than standard web protocols (http, xml based)

Johm Massonn (Ericsson): Privacy in the Web of Things - lot of data on WoT will not be protected - home sensors give detailed information on how we live -


Interaction Models for the Web of Things, Erik Wilde

<kaz> Erik's slides

<marie> scribe: marie

<scribe> scribenick: marie

Breakout Session #1

@@: service descriptions should focus on developers needs - simple APIs are prefered

scribe: also foucs on relevant market places
... let's focus on reviews, payments details, etc.

[minutes of this breakout session were already sent to dsr]

Breakout session #2 - Interaction Models for the Web of Things

Erik Wilde: may be those models are able enough, and cover most if the usgae scenarii

scribe: within Siemens, scenrios include home and office contexts, acting as hubs where we would gave a transversal interaction services
... exposing the services in a uniform way
... all is in fluc anyway. The social Web WG is starting, so I'd recommend to check what's happening in that space
... also, let's explore how to reach out to this social web group from with what was discussed during this wsp

[this breakout session minutes are also avail.]

Breakout session #3 - Business models for the WoT

Alession Gugliotta: standards for APis, standards for hardware developers, as well, as reuse of existing standards (security, etc.)

scribe: we also discussed revenue models and suggested some

[minutes to be sent]

Breakout session #3 - The WoT friendly label

Dominique Guinard: (has slides)

scribe: we felt that coming up with a label was important, since becoming an upstream topic
... WoT should be the app layer of the IoT
... we looked at several freindliness layers ... may be come up with a set of guidelines
... what should be friendly? devices, clouds, etc. May be have a WoT friednly stamp
... what would a bare minimum to have a friendly device?
... we also talked about different protocols
... and that there should be support to all of those and that bridges between those should be avail.
... finally, the findability layer
... how do I find my device on the Web
... reuse what's already in place, such as Web security
... thinking to create sth like schema.org for the IoT
... encapsulated Web objets including the data and the metadata

[breakout session notes to be sent later]

Breakout session #4 - Privacy

@@: around 12-13 people. Strong agreemnt that this is a strong issue, but also an hard issue to solve

scribe: different regulations in dieffrent countries do exist
... need o find breaches to pursue legal actions
... some people believe that regulations are the only way
... others say that regulators miss the point
... right to revoke data access is a minimum need
... group thought that W3C shoudl play a role to draft a policy data

(notes to be mailed out)

Panel session on core technologies - moderated by Jörg Heuer, Siemens

Jörg: what technologies and what enable these technologies

[introducing panelists]

Charampalos Doukas, CreateNet

CD: purpose of my talk is how to bridge Web to the IoT
... it's up to dev. and siftware producers to select the best
... about W3C contribution: it can identify solutions to the app level
... the COMPOSE approach is to use JSON and meta tags
... to describe REST inside messages... W3C to standradize a schema?
... I'll show a demo tomorrow - check it out
... links of interest: glue.things, cloud foundry, Node Red... if you want to play with our platform, go to mobiledemo.compose-project.eu

Matthias Kovatsch, ETH Zurich

MK: "Let the WoT begin at the device level" is mysimple statement
... device interoperability
... usability of the Web to create new and more complex apps
... creating value, optimizing processes are the goals

Markus Isomäki, Nokia

MI: my statement is "Interaction between Web clients and 'Constrained' Smart objects"
... withing the owp, we are missing the ability to communicate with these devices
... the CoAP protocol - an API for that would be useful

Milan Patel, Huawei

MP: items of considerations in order to choose core techs for the WoT
... message today, that let's reuse as much as we can
... every device/sensor out there should be LTE connected
... having gateways allows protocol transmissions, and a certain level of security as well
... providing to dev. effective toolkits
... consistency of data sets across use cases
... privacy should not be a barrier to innovation

Charles Eckel, Cisco

CE: real time communications, video conferencing, etc. is my background/experience
... comm between apps and network, with additionnal devices - how are these are going to exist and operate in an optimal way
... network providers need to support a wide range of apps.
... apps to interact directly with the network
... the network should give back the information on the devices' constraints

Jörg: quite different perceptions on the panel topic

<ph> scribenick: ph

<marie> scribe: Philipp

Mathias: ...

<dsr> Charles: Money will be a big driver of what gets deployed.

Charles: web of things is presentation of information, understanding it - money will be a driver - desire of people to share information could be a driver -

Milan: most applications have human end consumer from brainstorming i've done today and earlier - strong driver for wot is power of mashing together data from different sources

and add value - health, energy, disability, sports, .... - there will also be devices interacting with each other

Markus: don't see open standards in industry right now - many silos - web just used for representation of data to the user - interaction with the device pre-determined by the vendor

more opennes driven by developers and potential to mash up data coming from different vendors - something single vendor silos are not able to do

Matthias: a lot of people already noticed silos don't work - coap was designed to connect devices using web patterns - oma has device management standards - created lightweight m2m standard that builds on top of coap -

there is also ipso alliance - smart objects connected through ip

Joerg: is there a definition of web of things? is part of internet of things? do we have same understanding? What is difference between IoT?

Matthias: web and internet relation is same as wot iot relation - web is application layer of internet - wot plays same role for iot

Markus: pretty good definition - there's a bit of utopia in there, but I think it's reachable looking at the success of the web

Milan: web of things is what will add value to the internet of things

Charles: web of things will bring iot to vast amount of people - just like web did for computer usage - for everyone

Charalampos: web of things is the user side of internet of things - web of things is using technologies that browsers can communicate with

Joerg: would like to open it up a little - we observed that APIs are needed to get developers interested - others said to focus on protocols - Markus, where do you see balance between API and protocol technology?

Markus: protocols are just the necessary tools - Web has worked well without introducing many new protocols - there is webrtc standardisation that required new comm protocol (UDP, RTP)

if we want to enable local comm with constrained devices we may need some new protocols in the web platform

we have xmlhttpreqest API right now - maybe we need sth new

or a higher level API where you don't need to think about protocol

but additions should be minimal

COAP is on that path in my opinion

Wolfgang Daust(?): regarding protocols - things means something physical - it has not only a sensor, but also an actuator - completely different from a human being - different response times

the new thing is relationship to time - it matters for things

Charles: i agree with that - that's why I mentioned network constraints

you may want to know what is realtime and what is just bulk sensor data

Charalampos: not a matter of a new protocol - it's not web on things, but of things - some decisions can be done locally

Joerg: on the application layer, what are the technologies we see as part of web of things - we talked about cross-vertical integration - what technologies are available to descsribe information we are exchanging, and to describe the services

"please no new protocols" - are we already there?

Charalampos: there are efforts to describe sensors semantically - how to describe services is also important so that they can be reused easily

Matthias: we can learn from the web - many services that are easy to discover - use linking - many ways to describe services - microformants, schema.org, ... something is emerging - we still need some experience with wot, but am confident that we can pick one of the existing approaches
...: we talked a lot about data - but things can have logic embedded as well - but they are constrained - constraints are a difference between web of things and web of services

requires rule engines embedded into these things

Charles: one of the presentations this morning discussed this well - data will be passed into the cloud to do crunching - if thing has more capacity, can do computation ittself

Charalampos: ...

Matthias: question is where you put your RESTful API - ... - can push it into the cloud

to save money

on device

Johannes: is electrical vehicle a thing, or too big? what is a thing in WoT?

Matthias: in Web, you link resources together - resources can be composed - you can make it as big as you want, but also drill down do lowest sensor

Markus: similar to question: what is a resource?

Joerg: ... what are your expectations towards W3C?

Matthias: core components that made the web evolve - html - missing something like html for things - good representations, internet media types - should be reusable for many domains - not a model for each domain

should have core data model that can be reused and specialised

protocols we need more - also api - today web defined by what's in a browser - would like have a coap api, or even better a REST API

Markus: web sockets have brought asynch comm - webrtc bringing in p2p and realtime - communication with constrained devices still missing - COAP API - device apis there has been a lot of work already

there will be more of that

how do you describe services and resources third area of work

Milan: identifying technologies that are applicable to wot and iot w3c should do - w3c has re-examined technologies whether they fit mobile - should ask question again - are technologies wot friendly - how to make subtle changes so we can accomodate it given market is massive

Charles: I hope W3C won't do a whole bunch of new protocols - W3C should focus on APIs, data models

also something to replace inefficient polling - some sort of notification

Charalampos: would like to see more events like this, more open to developers - need to listen to their needs


Joerg: fog has cleared - look out of the window - there is some light

but still a little foggy

important to continue to talk to each other and continue conversations we started also after tomorrow evening

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.138 (CVS log)
$Date: 2014-06-25 16:40:42 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.138  of Date: 2013-04-25 13:59:11  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/Tokio/Tokyo/
FAILED: s/\. now/.. now/
Succeeded: s/slosts/slots/
Succeeded: s/conected/connected/
Succeeded: s/stic/stick/
Succeeded: s/sensors/sensors)/
Succeeded: s/,,./.../
Succeeded: s/demo/demo on localhost/
Succeeded: s/@@@/Andreas Harth, KIT/
Succeeded: s/@@@/Andreas Harth, KIT/
Succeeded: s/###/Johannes Hund, Siemens/
Succeeded: s/@@1/Jon Nordby, The Grid/
Succeeded: s/future/future (in 2020)/
Succeeded: s/Dave/Dave Conway-Jones, IBM/
Succeeded: s/dave:/Dave Conway-Jones/
Succeeded: s/Jones/Jones:/
Succeeded: s/Q (??)/Q (Andreas Harth)/
Succeeded: s/buiy/buy/
Succeeded: s/Appllicaiton/Application/
Succeeded: s/1é/12/
Succeeded: s/iOt/IoT/
Succeeded: s/ap/app/
Succeeded: s/is a /is my/
Succeeded: s/CUAP/CoAP/
Succeeded: s/what/that/
Found Scribe: fsasaki
Inferring ScribeNick: fsasaki
Found ScribeNick: kaz
Found Scribe: Marie
Inferring ScribeNick: marie
Found ScribeNick: marie
Found Scribe: ph
Inferring ScribeNick: ph
Found ScribeNick: ph
Found Scribe: marie
Inferring ScribeNick: marie
Found ScribeNick: marie
Found ScribeNick: ph
Found Scribe: Philipp
Scribes: fsasaki, Marie, ph, Philipp
ScribeNicks: fsasaki, kaz, marie, ph
Present: many many people
Agenda: http://www.w3.org/2014/02/wot/agenda.html
Got date from IRC log name: 25 Jun 2014
Guessing minutes URL: http://www.w3.org/2014/06/25-wot-minutes.html
People with action items: 

[End of scribe.perl diagnostic output]