Re: [Cfrg] NSA re-org and its impact

On Tue, Feb 2, 2016 at 7:54 PM, Tony Arcieri <bascule@gmail.com> wrote:
> On Tue, Feb 2, 2016 at 12:14 PM, Salz, Rich <rsalz@akamai.com> wrote:
>>
>> The NSA is re-organizing to merge its signals intelligence (attack) and
>> information assurance (protect) into one unit. The NSA is, by US Law, the
>> official advisor to NIST on cryptography. NIST has a pretty admirable
track
>> record of crypto (exceptions being mostly when they were misled by their
>> official expert). Things may change now. Or not. YMMV.
>
>
> At a time when NIST needs to restore trust, it would seem rather unwise
for
> them to accept any unjustified parameters from the NSA like they did with
> Dual_EC_DRBG (and hopefully they prefer well-scrutinized, widely trusted
> standards such as the CFRG curves)
>
> Dual_EC_DRBG was a debacle, but short of the NSA making massive advances
> over the public sector in cryptography and therefore being able to hide a
> backdoor in plain sight it seems like the sort of trick they can only pull
> once...

Like putting MD5 into widespread deployment after Dobbartin's 1996 paper?
Or Dual EC post 2007? Or continuing to widely use TLS 1.0 post Bard's 2004
attack, and Bodo Mueller's cbc-attacks.txt? Or making IKEv2 and IKEv1
sufficiently baroque that the only widely deployable mode was easily
decryptable, then ensuring that VPN makers would tell their clients to
enable it. We've know about every public NSA attack on crypto years before
it happened.

The basic fact is people know things about crypto, and will tell you what
to do. And you do them, or you suffer the consequences. And to the extent
the IETF can't adapt to that, we need to take the letter E out and replace
it with B.

>
> This is clearly a bad development, but hopefully NIST learned its lesson
the
> last time around.
>
> --
> Tony Arcieri
>
> _______________________________________________
> Cfrg mailing list
> Cfrg@irtf.org
> https://www.irtf.org/mailman/listinfo/cfrg
>

-- 
"Man is born free, but everywhere he is in chains".
--Rousseau.