Trust Registry TF — Darrell O'Donnell
ACDC TF — Samuel Smith Phil Feairheller
- Last week the meeting had a presentation from Kevin Dean from GS1.
- GS1 has many different needs for chained credentials: measurements, authorization, identifier delegation, etc.
- Part 2 of that meeting will be next week to map GS1's use cases into ACDC credential chaining.
- The GLEIF use case is now going into a public beta.
- A sandbox has been set up with a cohort going through a set of POCs.
- Scheduled to go into production in early 2022.
- They are starting to work on IETF specs.
- One of the first is SAIDs (Self-Addressing Identifiers): https://datatracker.ietf.org/doc/draft-ssmith-said/. Comments sought.
- Another is the CESR specification: https://github.com/WebOfTrust/ietf-cesr. CESR supports composability: concatenation of any type of cryptographic primitives in either binary or text format. It is truly a streaming protocol, so you can compose streams of any type and convert between text and binary, so you don't have to choose between "JSON or CBOR".
- Darrell said it would be very helpful to have Sam present to the Technology Architecture TF so we can figure out where CESR and KERI fit in the stack.
- Sam also clarified that ACDC uses JSON Schema, so you can build anything on top of it.
- We then discussed privacy options. Selective disclosure can be built on top of ACDC credentials, however Sam has also developed a hidden attribute capability.
- This allows an ACDC credential to contain only hashes of data, which the holder can then selectively disclose only the hashes desired.
- For GLEIF, they have developed a transaction event log (TEL) for transactions that are fully public (the chain of GLEIF credentials). But GLEIF does have one type of credential containing personal data.
- So Sam has developed an XOR blinded registry to use for cryptographic accumulators.
- Sam and Darrell agreed that the easier we can make it to adopt, the better.
- Sam agreed there are cases that need ZKP, but there are others where using blinded attributes is sufficient.
- Sam also pointed out that ZKPs are actually non-productive when a transaction needs to be non-repudiable, such as many regulatory use cases.
- AnonCreds is non-repudiable, but thus it is not true ZKP.
- ACTION: Darrell O'Donnell to attend the next ACDC Task Force meeting (Monday Dec 6th).
- ACTION: Drummond Reed to send an invite toSamuel Smith to attend the Technology Architecture Task Force meeting this Thursday Dec 2 — either 8AM MT or 2PM MT.
Design Principles TF — Drummond Reed
- The Design Principles for the ToIP Stack V1 has been approved by the ToIP Steering Committee as a ToIP Approved Deliverable. We are just awaiting final graphics to finish publishing it.
Technology Architecture TF — Drummond Reed
- This TF is steaming along. It has concluded that more than one ToIP protocol stack diagram will be needed.
- Two more meetings are planned this week and next week (Drummond is gone the week after that).