W3C

– DRAFT –
WoT Security

01 August 2022

Attendees

Present
Jan_Romann, Jiye_Park, Kaz_Ashimura, Michael_McCool, Tomoaki_Mizushima
Regrets
-
Chair
McCool
Scribe
kaz

Meeting minutes

Minutes

July-18

approved

Issues

TD Issue 1497

TD Issue 1497 - Identifiers don't seem to rotate enough

McCool: still need confirmation to close this
… (adds a label "Propose closing")
… can be closed in the Aug 3 TD call

McCool's comments

Discovery Issue 303

Discovery Issue 303 - Personal devices and public/private TDDs

McCool: already closed
… (adds some more comments)

McCool's comments

Wide reviews

TAG review

design reviews issue 736 - Web of Things (WoT) Architecture 1.1

kaz: I was suggesting we have a joint meeting with TAG
… probably we don't need to wait until TPAC but have that meeting this month

McCool: good idea
… have you or Sebastian contacted them?

kaz: no, not from me

McCool: agree, the sooner, the better
… but we still need to fix our own action items too

TD Issue 1635 - Adjust Policy-Like Assertions

Security Issue 208 - Remove References to "Security Best Practices"

Security Issue 209 - Update "Security and Privacy Guidelines" prior to PR of other deliverables

McCool: still need to work on policy-like assertions for TD (Issue 1635 above)
… also for Discovery and Architecture as well

TD assertions

privacy-mutable-id-ownership

McCool: (skims the assertions)

Jiye: it's more deployment information
… could have clearer description, once onboarding/offloading is clearly specified

McCool: right
… in that case, what can we do now?

Jiye: we can remove the assertion itself. maybe?

McCool: opt1. deleting the assertion
… opt2. refine the text
… opt3. clarify it's a policy
… think opt 3 would be the best
… and the question is how to describe that
… (adds some comments about possible rewording)

privacy-td-pii

McCool: need rewording

jr: making the capital SHOULD lower case?

McCool: (generates a proposed text)
… "As a matter of policy, it is suggested that THing Descriptions associated with a personal device be treated as if they contained personally identifiable information, even if this information is not explicit"

jr: sounds good

Jiye: fine

sec-inj-sanitize

McCool: similar rewording here as well
… "As a matter of policy, it is suggested that strings sourced from TDs either be sanitized using a carefully betted HTML sanitizer that diabls any markup or be inserted into an HTML template using DOM node manipulation APIs that will escape any markup."

security-update-contexts

McCool: two assertions here

Constrained implementations SHOULD use statically managed and vetted
versions of their supported context extensions. Constrained
implementations SHOULD NOT follow links to remote contexts.

McCool: then

Supported context extensions on constrained implementations MAY be
managed through secure software update mechanisms.

McCool: we could just delete it or make it a feature-at-risk

Jiye: opt 1 is fine for this

kaz: just to make sure, the first two assertions within one sentence remain
… while the last one "Supported..." to be removed
… right?

McCool: right
… suggest option 1 (delete it) for security-update-contexts (Supported context extensions...)
… but fold the idea of "secure" updates into an earlier assertion: "Constrained implementations..."

=> "Constrained implementations SHOULD use vetted versions of their supported context extensios managed statically or as part of a secure update process."

McCool: what do you think?

(no objections)

McCool: will generate a PR for that in time for the TD call on Aug 3

[adjourned]

Minutes manually created (not a transcript), formatted by scribe.perl version 192 (Tue Jun 28 16:55:30 2022 UTC).