10D/ Directories in Distributed Identity
Directories in Distributed Identity
Wednesday 10D
Conveners: Sam Curren, Ken Ebert, Suresh Batchu, Kiran Addepalli
Notes-taker(s): Kiran Addepalli [kiran@digitaltrust.net]
Discussion notes, key understandings, outstanding questions, observations, and, if appropriate to this discussion: action items, next steps:
Slide Deck: https://docs.google.com/presentation/d/1YjTJK1Zq8Z5iRmo3cn321EUu4fjwpCci/edit#slide=id.p1s
ADIA site: https://adiassociation.org
Does the Directory support a pairwise DID - The directory supports exchange between two parties. The directory enables the interaction.
Would including a DID in an LDAP be a good way to implement the directory?
Much of the social web doesn’t require KYC. Trusted Issuer = Trusted by the Directory. We don’t prohibit the model where self-asserted identities are not left out. Ex: Email provider can be a trusted issuer. Some services like Instagram are getting good at providing ads that are relevant.
Directory is responsible for the DID and then it is up to the parties to decide what information to share.
Will Alice be notified by the directory when one of her DIDs has been shared with Bob — or only if/when Bob reaches out to her via the DID obtained from the directory? - Yes
I am just wondering if there is probably a need for syncing in a decentralized manner. E.g. I am registering a Service at a specific Directory Service (because I know exactly that one) but I want to be distributed (in a verified way) that the service will be available on all other directories which somebody else want to trust.
How would customer integrate with the ecosystem
We have the specification coming up in june. The technology working group meets on thursdays. Please contact jason@digitialtrust.net for further details.
The specification also aims to address some of the protocols to interact with the directories.
Will the Directory analyze/store query results to optimize future query results or would that violate the Directory’s privacy model? - we are focussing on the metadata lookups for faster lookups.
I think there is room for zero-knowledge service type directories that don’t know about the data they process, as well as transparent ones that do query optimization and analytics
Dan Robertson - On an earlier topic, even when hashed search values are used, a client could potentially do fuzzy matching with a local list of synonyms, where a user inputting "dan" would also automatically get results for "daniel" and "danny".