On 03/27/18 05:43, Paul Hoffman wrote: > On 26 Mar 2018, at 17:30, Michael Sinatra wrote: > >> I am a bit uncomfortable with the document's disrecommendation of SHA384 >> and ECDSAP384SHA384. The main reason for this is that for crypto >> recommendations here in the USG, > > Note that those are for encryption, where they want to keep some things > secret for 40 or more years. DNSSEC is an authentication mechanism. If > it takes two decades to break an P256 key that is used for encryption, > the attacker gets much (much!) more value than breaking one that is used > for authentication.
Agree, and I would even add that the limited signature validity interval of DNSSEC makes ECDSAP256SHA256 more than "good enough" for DNSSEC. However, my motivation is based on your next point... > The fact that the NSA document is unclear on this point has been causing > problems for many people, including for people inside the NSA. My goal is to basically avoid confusion and just tell people to use the strongest algorithm they can reasonably use. I.e. follow the CNSA recommendations and don't spend a lot of time thinking about the application. I'll try to come up with some wording for the draft (i.e. the sentence that I was uncomfortable with) that captures the notion that P-384/SHA384 doesn't add much in the way of security, but that there's nothing wrong with "crypto maximalism" in this case, and see what the WG thinks. michael _______________________________________________ DNSOP mailing list [email protected] https://www.ietf.org/mailman/listinfo/dnsop
