5 ms·
Did literally anyone not see this coming? We all know the government can’t hold on to keys. I fear this will be used as an excuse to make the passport system e
by mwint 5y ago
Did literally anyone not see this coming? We all know the government can’t hold on to keys.
I fear this will be used as an excuse to make the passport system even more centralized.
- Bancakes 5y agoWhy the downvotes? For sure, this will just have us install (or have preinstalled) DRM-locked proprietary apps on our phones. Ugh.
- ethanbond 5y agoDownvotes because governments regularly hold secrets for years, decades, sometimes centuries. But yeah yeah, you can point to more than a few failures in the massive sample pool of “the government” so “government can’t hold keys.”
- gpderetta 5y agoAs opposed to private entities that are known to be able to keep keys secret all the time?
- switch007 5y agoThey didn’t suggest that at all.
- cm2187 5y agoDon't know if this is whst the OP has in mind, but the first thing that comes to my mind is the backdoor that many states want to impose on all encrypted communications. That would create a single point of attack, so even assuming competence, it makes the system significantly weaker. And this episode if confirmed makes competence a strong assumption.
- gpderetta 5y agoIt very much points to incompetence. I just don't think governments are specifically more incompetent than private entities in this case.
- Arnt 5y ago78 governments have ICAO private keys, and most have done so for >10 years now. If what "we all know" is true, then you should be able to find a leaked key easily. Try googling. Or you might fall back to claiming that while governments evidently can hold on to ICAO private keys, they can't hold on to this other kind of private key, because...
- Scoundreller 5y agoMight be a matter of issuing bodies per country. Most countries have 1 passport issuing authority that’s used to thinking about security 1st, while they can have many covid certificate issuers. And many clinical authorities are not security 1st.
- Arnt 5y agoThat's one option. The other is to regard the initial statement (that "We all know the government can’t hold on to keys") is false. Pick a test you believe to distinguish between the two. Say, whether the access keys for at least one database in a country that has >100 has leaked. Then google to see whether you can find such a leak.
- JumpCrisscross 5y ago> Did literally anyone not see this coming? The system's designers did, which is why key revocation is built into the system. The practical effects of this leak will be the people who refused the app and don't read the news will be surprised when their paper certificates are rejected.
- alberts00 5y agoAs far as I'm aware there was no technical solution for key revocation when the EU Covid Certificate was first launched in July. The only possibility I saw for revocation was to revoke the whole CA, instead of e.g CRL check. Can you elaborate what makes you think that key revocation is built into the system?
- ghostpepper 5y agoI'm a bit confused by this also. My understanding was that each node in a certificate tree/list consists of a key pair (public/private), and the entity metadata which needs to be verified. eg. the root CA has a private key that it uses to sign CAs, and those CAs have their own private key that it used to sign individual leaf node certificates. Meanwhile, revocations are distributed as a separate CRL - certificate revocation list - which contains a list of certificates whose signees are no longer to be trusted. I'm not very clear on how this process works, but in any case I don't think keys can be revoked.