Meeting Schedule
- Bi-Weekly at 8:00-9:00 am PDT / 11:00-12:00 am EDT / 15:00 – 16:00 UTC / 17:00 - 16:00 CEST
Attendees
- Eric Drury
- Carly
- Feng Hou
- Charles Lanahan
- Drummond Reed
- Neil Thomson
- Scott Whitmire
- Jordan Evans
- Darrell O'Donnell
- Steve Magennis
- Judith Fleenor
- Stephen Curran
Main Goal of this Meeting
We're using this week's EFWG meeting to give the floor to anyone who'd like to give an update on IIW.
Agenda Items and Notes (including all relevant links)
| Time | Agenda Item | Lead | Notes |
| 5 min |
| Chairs |
|
| 1 min | Announcements | TF Leads | News or events of interest to EFWG members: |
| 5 min | Review of action items from the previous meeting | Chairs | Bhutan NDI Case Study update |
| 40 min | IIW Recap | All | Sessions or content that we'd be interested in hearing about:
|
| 5 min |
| Chairs |
Recording
Notes
AI notes - coming
TDW - Trust DID Web is a new DID method but doesn't get DID doc from HTTP location but instead you get a log of all the entries of the changes of the DID doc. Every line is tied to the previous via a hash of the entry. The controller determined according to DID spec who much sign every transition. Very similar to DID: webs, but the difference is the transition state is the DID doc. Can you pre-rotation of keys. Next step, take the spec to a task force to evolve the specification. This DID has portability that lets you move the location of the DID, this changes the DID but the SCID (self certifying identifier) plus the history stays the same. Combine with high assurance DNS as with DID:webs will work the same. Long term storage - 30yr+, what kind of archival storage for this information?
DID:web is becoming well used but it really lasts history verifiability and this specification (TDW) adds ledger based features without a ledger but still fits very well with DID:web.
Comparison to DID:webs - DID:webs is KERI focused with a side of DID. But DID:tdw is DID focused and has less complexity compared to KERI and DID:webs.
DID:webs - uses KERI to take you state to state (using KERI and then producing a DID document out of that), you have to go further to generate a DID doc, requires more complexity to fit within the DID standard specification. The implication is that DID:tdw is easier from an adoption standpoint.
did:tdw Specification (rendered): https://bcgov.github.io/trustdidweb/
did:tdw Specification (repository): https://github.com/bcgov/trustdidweb
Presentation at IIW -- the details start at slide 11: https://docs.google.com/presentation/d/1PHo16asyceRiNKN7UkV8BSmtWtN6Wp3A6_9MV0IQ2jg/edit?usp=sharing
Typescript Implementation: https://github.com/bcgov/trustdidweb-ts
Python Implementation: https://github.com/bcgov/trustdidweb-py
Compared to DID:webs there are some risks compared to KERI but it is more secure than DID:web.
Is there a migration path from DID:tdw to DID:webs - they are totally different and it would have challenges.
Witness, watcher features of KERI could be added to DID:tdw but it doesn't belong in the specification but may appear in the implementation guide.
Apple session - prof of concept across android and IOS. Asked to provide a credential on a website but your credential is stored on your mobile device and they create a tunnel to establish a connection between the browser and the mobile device. The mobile OS can then ask different wallets for the needed credential which can be passed back to the browser where it can be shared. Interoperability across the two, use of protocols (uses the method that passkeys uses to communicate, it doesn't use passkeys: https://fidoalliance.org/specs/fido-v2.1-ps-20210615/fido-client-to-authenticator-protocol-v2.1-ps-20210615.html). They make no standardization calls or opinions on standards. Multiple wallets on a device - Apple may allow that on their devices.
Trust registries - about 7 trust registry initiatives that are being worked on, they were outlined and how they operate. Checq, ToIP, OID federation, trust establishment with DIF)
EBSI Trust Chains: This standard tracks “Verifiable Accreditations” and is also used by cheqd. It involves a governing authority for the ecosystem with a DID on a blockchain, tracking DIDs authorized for specific actions.
Trust over IP Trust Registry Protocol v2: Version 2 is under implementor’s review as of April 2024. See this ToIP blog post for a full description. It offers a RESTful API with a query API standardizing how to query which entities are authorized to do what in which context.
OpenID Federation: This standard, particularly OpenID Federation 1.0, is already used in systems worldwide, including university networks and Brazil’s open banking. It allows each entity to provide trust lists, including common trust anchors with other lists.
Credential Trust Establishment 1.0: This standard, part of the DIF Trust Establishment specification, is a data model rather than a protocol or interaction model. It involves creating a document and hosting it behind a URI, with no centralization. It allows roles for each participant and is complementary to VC-based decentralized trust.
There was also brief discussion of two others: TRAIN, from the Fraunhofer Institute, and the W3C Verified Issuer/Verified Verifier model.
An elegant solution to work X.509 into DIDs - use alternative names field of the X.509 certificate to include the DID reference. No need to create a separate DID method to work with X.509.
Eric Scouten at Adobe also co-chairs the X.509 VID Task Force at ToIP (meets every other week, Thursdays 8:30AM PT—I attend most meetings). The goal is to build a bridge between X.509 certs and decentralized identifiers (DIDs/VIDs) so that an ecosystem or an issuer does not have to choose one or the other. After researching all the options, it has become clear the best one is also the easiest: just publish a DID/VID in the Subject Alternative Name field of the X.509 cert. That makes it easy to go from the cert to the DID/VID. (To go in the other direction—from the DID document to the X.509 cert—there are several options, including putting a specific service endpoint type in the DID document.)
To do a sanity check with the IIW community on this design and on the value of an X.509-to-DID/VID bridge, on Thursday Eric and I called a session together with WebTrust auditor Scott Perry and BC Gov architecture Stephen Curran (who had already given his Last Great DID Method session). We had a number of X.509 savvy architects and developers attend, plus a woman from Digicert who used to work on X.509 at Adobe.