W3C

- DRAFT -

WoT PlugFest/Test

03 Oct 2018

Attendees

Present
Kaz_Ashimura, Michael_McCool, Takeshi_Sano, Ege_Korkan, Tomoaki_Mizushima, Kunihiko_Toumura, Michael_Lagally, Ryuichi_Matsukura, Takeshi_Yamada, Toru_Kawaguchi
Regrets
Chair
McCool
Scribe
egekorkan

Contents


<kaz> scribenick: egekorkan

Agenda

<kaz> McCool: online plugfest report and lyon planning

Fujitsu report

Matsukura: Sano-san will give Fujitsu's report

<sano> https://github.com/w3c/wot/blob/master/plugfest/2018-sept-online/result-fujitsu.md

Sano: as fujitsu we will provide td directory and proxy
... we provide authentification, basic auth.

McCool: did you add https?

Sano: no only http

McCool: so you added basic oath but it was only with http

Sano: this figure is about sequence diagrams between directory, servient and proxy

Kaz: could you make the page bigger

Sano: application could connect to the devices via proxy, directory
... but some devices couldn't connect to the proxy, oracle reported that
... now we fixed the bug

McCool: so i see that you tested bearer and oath
... for the intel devices, since you tested http and https, they were using reverse proxy

Sano: we connected on wlan device

McCool: so you didnt try the online devices?

Sano: no

Kaz: different color for the lines, the ovals or the background to distinguish the VPN world from the Internet world?

Lagally: maybe use colon instead of percent?

McCool: td should be able to deal with that

Sano: but proxy should do this encode

McCool: speaking of validity, takeshi san are you done?

Kaz: ml, which would be better, considering the BluePump from Oracle as a device? or part of the Oracle IoT Cloud Services?

Lagally: what I understood, is the blue pump (connected to the cloud)?
... we can simulate multitude of devices

McCool: this is not the overall system. This figure shows the work of Fujitsu. Intel has also more devices

Kaz: this is what you see from your point of view right?

Sano: yes

Lagally: It is a standalone device, open to the internet protected by basic auth at the moment

<inserted> Kaz: so Lagally, this diagram having the Oracle BluePump as a blue bubble is fine from your viewpoint as well. Right?

<inserted> Lagally: yes

McCool: intel has a similar architecture to oracle

Validation

McCool: i want to talk about validation
... we used the validation tool

Ege: which tool

McCool: we used the playground but who has used the network interface devices

<kaz> McCool: we changed the policy for security within the TD spec

Ege: i have tried to use that tool for the previous plugfest

<kaz> McCool: security is mandatory at the top level, and optional at the other levels

Ege: but I havent used it for this one

McCool: we can have a private meeting to use the vpn to test the validation tool on intel devices

Ege: yes that would work, we can do it next week

Kaz: will we put it in the services for Lyon?

McCool: I have it here in this
... used but had a difficult to understand error

<kaz> Kaz: can we use the playground validator as a web api?

<kaz> McCool: there is a command line version as well

Ege: but I use tool to validate so I cannot control the error messages
... there is a scripted version but not as a service
... we have also talked about merging the UIs of playground and directory

McCool: but it would be better to do this as a service so that we can automate it

<McCool> https://github.com/thingweb/thingweb-playground

McCool: I can find it here on github
... so I can use it locally and then register only if it validates
... generally speaking, input validation is a good thing

Lyon preparation

McCool: we should organize our organisation for the lyon plugfest
... we will be there in person
... I started using .local URLs in my TDs
... if people have IP addresses, we can use it like this
... so that transition (to ?) is easier
... this thing (directory) has no security
... so we should add it that evil people cannot abuse it
... koster had an hacker who did something to his node red service that had no security
... if you leave something online with no security, someone will find it
... listing the tools not as services

Lagally: current list is very unstructured, some things are not available. we can structure it so that what is said available is always available and things that are not always available are structured
... I created a PR
... about scenarios

McCool: (merges the PR)
... I think target features belong in the readme

Kaz: we should also list onlineservients?

McCool: services are thing that are not described/found in the td directory

<inserted> Kaz: btw, I've started to wonder the relationship between "Services" and "Things", since there could be "virtual Things" as well.

McCool: service has architectural purposes

Ege: the network interface tester is a service but is implemented as a thing and has a TD
... I didn't how to make a UI for it and so on, so I implemented it as a Thing such that it has a TD that everyone can understand

McCool: I see
... we can move these already completed things into available services structure

Ege: you can also put the network interface tester into work in progress

Lagally: listing already available services that can be used even before the plugfest

McCool: the services listed under available are already running at the moment
... (organizes the services and things into categories)

Lagally: you can put the ORacle cloud into infrastructure

McCool: we can add more proxies here that I don't know of
... if others have suggestions we can do it here

<kaz> Kaz: at some point, we should add possible applications as well

McCool: I don't think these belong in the readme of the plugfest
... regarding the wlan, i took down my devices today
... any comments?

Kaz: thank you very much to every body
... for attending the online plugfest

McCool: it was a lot of work but learned a lot

<kaz> [adjourned]

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version 1.154 (CVS log)
$Date: 2018/10/04 00:43:51 $