Re: Revocation and No Phone Home

On Mon, Jun 2, 2025 at 5:31 PM Pryvit NZ <kyle@pryvit.tech> wrote:
> 1. This is the issuer spoofing scenario. An issuer creates a bitstring list of sufficiently large size, but only includes 1 active credential index per bitstring list.

We highlight this scenario here (and suggest some mitigations):

https://www.w3.org/TR/vc-bitstring-status-list/#malicious-issuers-and-verifiers

> 2. This is the degradation scenario. A sufficiently large credential set is established per bitstring list, but the expiration period is vastly different per credential.

This is an excellent point and something we don't cover in the privacy
considerations section. I've created a new issue to track the concern
and add some privacy considerations text:

https://github.com/w3c/vc-bitstring-status-list/issues/206

We'll do this as soon as we publish the first draft of the Bitstring
Status List v1.1 specification (expected towards the end of the
summer).

> This seems like requiring OHTTPS would make for a good solution if the issuer hosts these revocation lists. At least then even if K-anonymity is 1 then the only inferential knowledge the issuer gets is time and frequency based.

Yes, agreed. We might also strongly urge issuers to use CDNs for their
revocation lists as an "easy first step" since many large issuers
already use CDN services for their websites.

> It would be nice to see some method that allows an honest holder or verifier to be able to act as a transparency watchdog to make sure K-anonymity never reaches 1, but I’m not entirely sure this is possible. Maybe someone on this mailing list has a clever idea of how to improve it further.

We say something about opt-in watchdog services here:

https://www.w3.org/TR/vc-bitstring-status-list/#malicious-issuers-and-verifiers

I'll also note that, due to the way Bitstring Status List is designed,
it can be fetched anonymously (via OHTTP) by the holder and delivered
directly to the verifier as a VC. I don't think this pattern is
largely implemented, but because the BSL is a VC itself, you can
deliver it in a peer-to-peer fashion and as long as the verifier is
comfortable with its "freshness" (VC is less than 24 hours, 28 hours,
7 days old), then even if the K-anonymity set is 1, the holder can
hand-deliver it to the verifier where the verifier doesn't need to
contact the issuer.

We might want to rank all of these ways to improve upon the privacy
characteristics of contacting the issuer for a revocation list, by
implementation burden and operation cost, in the BSL specification.
I've opened an issue to track that idea as well:

https://github.com/w3c/vc-bitstring-status-list/issues/207

-- manu

-- 
Manu Sporny - https://www.linkedin.com/in/manusporny/
Founder/CEO - Digital Bazaar, Inc.
https://www.digitalbazaar.com/

Received on Tuesday, 3 June 2025 13:23:46 UTC