Hacker News new | ask | show | jobs
by indolering 93 days ago
Bad arguments and FUD when it was being rolled out. Sysadmins also don't want to touch working infra code, you can see that with AWS lagging on IPv6.
1 comments

Who's the most reputable cryptographer you can think of who publicly supports DNSSEC? We'd like to interview them on SCW.
Can you not check the RFCs?
You know the funny thing about this is that I have talked, relatively recently, to one of the very few cryptographers who was an author on a DNSSEC standard, and they wouldn't work for the interview I want to do --- they're not sold enough on DNSSEC anymore.

The broader answer is: the relevant RFCs weren't authored by cryptography engineers. This was a major problem in the "old" IETF, before the cryptographers "took over" tls-wg and CFRG.

At any rate, the reason I asked in that particular place on the thread was that the preceding comment was attempting to draw a line between "sysadmins" who hate DNSSEC and the serious technologists who like it a lot.

You are going to complain that the key sizes are too small despite the guidelines being updated a long time ago. Then you will argue adoption of larger keys sizes is to low. Then you will argue that we should just not sign domain name authority delegation records at all (i.e. DNSSEC) and that we should abandon shoring up authenticated DNS because there is no adoption.

You have any cryptographers that are satisfied with unauthenticated name server checks?

Yes? Lots of them? But also: you didn't answer my question.
Okay, but after this I have to go back to work.

You got a point: 1k isn't great and of course mainstream cryptographers will advocate for higher. That doesn't change that it's still acceptable within the existing security model nor that better alternatives are available. The cryptographic strength of DNSSEC isn't a limiting factor that fatally dooms the whole project. We have to upgrade the crypto used in large-scale infrastructure all the time!

And yes, uptake of better crypto is poor but I find chicken-and-egg arguments disingenuous when coming from someone who zealously advocates to make it worse. Furthermore, your alternative is no signing of DNS records. Find me a cryptographer who thinks no PKI is a better alternative. I know DJB griped about DNSSEC when proposing DNSCurve, which protects the privacy of the payload but not the intergrity of the payload.

Is this a bot gone rogue? Parent asked for a person, and you are shadow-boxing with unasked questions.
The question was "can you find me some reputable cryptographers that support your position?" which is just ad hominem and should be ignored as such, except it does indicate that the person asking it doesn't have any better argument than ad hominem.
Sorry, but I asked who's the most reputable cryptographer you can think of who publicly supports DNSSEC? I asked because we'd like to interview them on SCW.
More rhetorical dunking instead of engaging with the substantive technical issues. I'm done.