26 ms·
DNSSEC root key signing ceremony postponed because they can't open a safe
- Foxboron 7y ago>during what was apparently a check on the system on Tuesday night – the day before the ceremony planned for 1300 PST (2100 UTC) Wednesday – IANA staff discovered that they couldn’t open one of the two safes. One of the locking mechanisms wouldn’t retract and so the safe stayed stubbornly shut. I clicked randomly around in the linked livestream from November and came across what seems to be them having problems with Safe 2. I'm amazed this is streamed and they are actively showing the written pads towards the camera. https://youtu.be/erfsFJsapAs?t=680 https://youtu.be/erfsFJsapAs?t=680
- _-___________-_ 7y agoYou mean the paperwork they're holding in their hands? Is there something secret on it?
- Foxboron 7y agoIt's for viewers to verify the work they are doing. I'm not implying they shouldn't.
- _-___________-_ 7y agoAh, thanks for the clarification. That's smart.
- loulouxiv 7y agoIs the HSM sacrificed when it has done its work ?
- erik_seaberg 7y agoI'm reminded of a story about Google's internal password vault outage, where restoring service required opening a safe whose combination was of course in the password vault.
- exikyut 7y agoTIL about this story. Naturally, trying to Google for further information about it isn't working :)
- brokenmachine 7y agoHow did they resolve it?
- erik_seaberg 7y agoThey expensed rubber mallets, which didn't work. They brought in a locksmith, which did.
- Spare_account 7y agoI'm surprised both of the key-signing locations are in the US. There should be more locations in other countries.
- saber6 7y agoIt’s not a USG power play. It just so happens that a large portion of internet engineers and architects live in the United States.
- tptacek 7y agoBut it's probably worth remembering that DNSSEC is a PKI over which the USG and its Five Eyes partnership has enormous, outsized influence; they have de facto custody over the most important TLDs on the Internet, a demonstrated willingness to intervene in DNS for policy goals, and, unlike a suborned CA, the TLDs cannot be revoked when the IC abuses them. "Ceremony" is the right word for IANA's handling of the root zone, which is a McGuffin in the Internet security model; the DNS roots won't be what the IC targets.
- saber6 7y ago> But it's probably worth remembering that DNSSEC is a PKI over which the USG and its Five Eyes partnership has enormous, outsized influence; they have de facto custody over the most important TLDs on the Internet, a demonstrated willingness to intervene in DNS for policy goals, and, unlike a suborned CA, the TLDs cannot be revoked when the IC abuses them. You're off base. DNSSEC is not any more nebulous than the existing structure and control mechanism with regards to domain names. It (mostly) is adding cryptographic under-pinnings to the responses. The control remains the same (registrar level).
- tptacek 7y agoI'm not off base, and the Five Eyes governments obviously control the most important TLDs, as anyone who was around during the piracy seizures knows. The control remains the same, yes: controlled by the USG.
- sneak 7y agoIs the safe going to be drilled on the livestream in front of the witnesses? Or are the key material devices and the HSM going to be accessed outside of the ceremony for potential tampering? Seems to me that exceptions and anomalies like this are the weak points of the system. If I wanted to steal these keys, I would make an effort to create anomalies like this where I might have physical opportunities to glitch the keycards and exfiltrate the data without the whole internet watching me.
- edf13 7y agoWith a tinfoil hat on - how do you know the livestream isn't pointed at some fake studio and the real safe hasn't already been tampered with?
- pas 7y agoYou never know, after all maybe your whole life just happens inside a big box (or a small one if you are a simulated/real brain in a vat). But if you accept that it's possible to gain some information, and to basically slowly build up a model of the real, so you can start trusting sources of information, eventually you can build a ceremony model and there are designated witnesses, and designated key-holders. And in theory press should regularly interview them to make sure they are trustful, mentally okay, they are not blackmailed, or otherwise tampered with, and what they say about the process, safe, stream, etc. So in the end it's all about your priors. If you have an almost unmovable belief about we're all just being lied to, then you might never (as in not in <100 years) update to the alternative belief.
- mrob 7y agoI don't know what the IANA is doing, but this is essentially the same problem as verifying election results, with the same solutions: https://en.wikipedia.org/wiki/Election_monitoring https://en.wikipedia.org/wiki/Election_monitoring
- exabrial 7y agoThey could sign a piece of data you provide (pseudo random proof) that they have no control over
- dependenttypes 7y agoDNSSEC you say? https://sockpuppet.org/blog/2015/01/15/against-dnssec/ https://sockpuppet.org/blog/2015/01/15/against-dnssec/
- egberts1 7y agoI stopped midway down the first paragraph.
- caymanjim 7y agoThat was written by a prolific anti-DNSSEC campaigner active here on HN. I don't understand the system well enough to critique it, but between that post and various comments on HN, I've stopped assuming that DNSSEC is a universal good thing. Before I recently hardened my own email server and had to jump through all the hoops to get DANE working, I thought that DNSSEC was actually DNS-over-TLS with some sort of chain of trust back to the authoritative DNS server, and was surprised to learn that it was none of those things. I'm still going to use it but I'm not going to evangelize it the way I would if it actually provided encryption and trust.
- alwillis 7y agoI'm still going to use it but I'm not going to evangelize it the way I would if it actually provided encryption and trust. It was never supposed to encrypt lookup; that was mostly a disingenuous strawman argument from the usual suspects. It does provide a chain of trust. By using the public key published in your zone, a validating resolver checks the digital signatures from the root to your zone. It’s no different than when a browser checks the digital signatures on a certificate that’s signed by an intermediate certificate that was signed by a root certificate. For authentication and trust purposes, every resource record in a DNSSEC-signed zone is cryptographically signed and validating resolvers--Unbound, Knot Resolver, BIND--confirm them.
- ElectronShak 7y ago>so that when you visit, say, theregister.co.uk, you eventually reach one of our servers at network address 104.18.235.86. More like you reach CloudFlare's CDN
- zhte415 7y agoI'm both reassured by this, but also concerned. Reminded of a former employer with the entire global data centre in one single location, quite a secure location, but not very secure against bad actors, that could take down banking across much of the world.
- BrandoElFollito 7y agoThis is the same case with travel reservations. It is surprisingly not resistant to internal attacks and would be one of the top 3 most severe disturbances in the world (I have a hard time imagining the other two - your banking example is one of them, if they are indeed that centralized)
- wbl 7y agoIf anyone used DNSSEC it would be a problem.
- microtherion 7y agoThe description of the ceremony is fascinating. Somewhat disappointed, though, in the lack of pre-modern flourishes. No goats sacrificed? No pentagrams drawn on the floor? No requirement for a full moon? Engineers don't have to take a lifelong vow of chastity? How do they ever expect a low-rent affair like this to work?
- londons_explore 7y agoDNSSEC would only be of relevance if records could cryptographically sign TLS certificates. Since no browser today can do that, DNSSEC is a dead technology.
- rhn_mk1 7y agoIt covers pretty much the same use case for email with DKIM.
- vegardx 7y agoI guess you could call all optional technology for dead then, but it's super useful for things like ssh fingerprints (SSHFP), and I imagine they're also useful for things like public keys for IPSec (IPSECKEY) and other types of certificates (CERT), like GPG public keys.
- alwillis 7y agoDNSSEC prevents certain kinds of attacks, like cache poisoning, on the DNS system and a little over 15% of internet hosts are behind resolvers that validate DNSSEC signatures when making DNS requests. You’re thinking about DANE’s TLSA records, which supports trusting certificates not signed by the CAs browsers already trust. DANE does require DNSSEC to work but that’s not what the article is about.
- tptacek 7y agoDNSSEC attempts to prevent certain kinds of attacks that have already been designed around in the Internet threat model; we have spent the last 30 years of Internet security design working towards an Internet where it doesn't matter if you can poison a DNS cache, because nothing trusts the DNS anymore. For most practical purposes, we are already there, and the trajectory to get all the way there (essentially: decisively and finally solving WebPKI certificate misissuance) is clear. DNSSEC doesn't matter; the parent commenter is right.
- wbl 7y agoThe definition of mississuance is based on DNS data so DNS does matter for CAs. Conceptually DNSSEC would help except no one uses it.
- masto 7y agoWhere is the lock picking lawyer when you need him? A little click on 3, some counter-rotation, and there you have it!