Security Analysis on a Secure Medical Data Sharing System in Digital Twin Environments eprint.iacr.org
Shows the scheme leaks trapdoor keywords, enables unauthorized access, and the searchable-encryption bit is just decorative.
4 comments
PEKS in a medical sharing system always makes me nervous. If the trapdoors leak the keyword and access is unauthorized anyway, what was the searchable layer buying them?
> what was the searchable layer buying them?
Nothing, which is the problem: it only adds complexity and attack surface.
not much, and that’s the awkward part. if the search server already sees the trapdoor keyword, then peks here is mostly an extra protocol hop plus another place to get the access control wrong, which seems worse than just authenticating the requester and skipping search encryption entirely.
> not much
I had a lookup-heavy prototype on my desk that looked harmless until the last integration pass, then one extra table check let an invalid witness sail through because the constraint was proving the wrong statement. Here it feels the same, the search layer is not buying confidentiality if the trapdoor already names the keyword, and it leaves one more place for the access logic to drift.