Hello dear Viktor Dukhovni.

Viktor Dukhovni wrote in
 <[email protected]>:
 |On Sun, Oct 25, 2026 at 02:13:03AM +0200, Steffen Nurpmeso wrote:
 |>|Its DNSKEY RRset
 |>|(including an ML-DSA-44 KSK, ZSK and RRSIG) arrives over TCP:
 |>|
 |>|    ;; Query time: 1395 msec
 |>|    ;; SERVER: 127.0.0.1#53(127.0.0.1) (TCP)
 |>|    ;; WHEN: Sat Oct 24 03:46:45 UTC 2026
 |>|    ;; MSG SIZE  rcvd: 6546
 |> 
 |> That is in my opinion just as strange as trying to use this for
 |> DKIM, .. but likely the TTL is even lower.
 |
 |It's TTL is 1 hour, and the deployment is I expect a deliberate
 |pilot exercise.

 |> I want to believe this is an operator error of someone who is
 |> either misinformed, or even hysteric.
 |
 |There's no reason to think so.  The algorithm is being tested at
 |Cloudflare (with domains that are thus far either private or in
 |any case off the DANE-survey's eTLD+1 radar).

I have a very strong opinion on that, really.
If it would be me i would instead reduce key lifetime so that
DNSSEC has all the options to switch to a better algorithm within
a year.  And ouch! that makes for plenty of better options in the
real life as i expect it, where even 2035 is likely happily living
without CrPQ, as you said.  So 2030, and by then i would really
expect new algorithms having entered even OpenSSL.

 |> Yes TCP retries forever, but the musl C library for example only
 |> ever implemented it in the not too distant past.  (Four years now
 |> that i look.)
 |
 |TCP support in the stub resolver is not what's important here, it

Well it is, MLDSA44 public key is way over 512 bytes, is it?
You need EDNS or TCP, even in the stub resolver, no?

 |is the system's validating recursive resolver that needs TCP, even
 |prior to the registration of algorithm 18:
 |
 |    https://www.iana.org/assignments/dns-sec-alg-numbers

That RECOMMEND[ED]s pretty short ones, mostly.  (It does not
recommend that much, talking about algorithm breakage.)

But .. you know, i wrote my caching stub resolver about the
quarter of a century ago, it was not even *that* DNSSEC by then
(and we did not support that out of the box), and i have forgotten
most of that, anyway.

I am super happy we now have DNS over secured transport, and am
totally thrilled that OpenSSL supports DTLSv1.3.  What options!!

(I would only wish i had more time for programming really, though
having to deal with hundreds of mouths and beaks each day is also
very fullfilling, is it.)

 |    https://datatracker.ietf.org/doc/html/draft-westerbaan-dnssec-mldsa-03/

I am not looking there.  It really is grazy, in my humble opinion.

 |> And then, it is "very expensive", at least in my world; there it
 |> is a fallback if the UDP did not work out.
 |> In a modern system this thus likely sends maybe even two UDP
 |> packets (IPv4 and IPv6), and then (hopefully only one) a TCP
 |> fallback retry.
 |
 |That's not how DNS works.  The 1st UDP query elicits a truncated
 |response (TC=1) (one RTT) and then the client switches to TCP,
 |(3-way handshake, query, response, which is 2-RTT in the absence
 |of TFO and/or a cached TCP connection for a domain that is queried
 |often).  When I suppress TCP with "+ignore", dig queries to my
 |local cache report:

I also use dnsmasq, i think most of local DNS caches by now.
That was not always the case!

 |    ;; global options: +cmd
 |    ;; Got answer:
 |    ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 36254
 |    ;; flags: qr tc rd ra ad; QUERY: 1, ANSWER: 0, AUTHORITY: 0, ADDITIONAL\
 |    : 1
 |
 |    ...
 |
 |    ;; Query time: 0 msec
 |    ;; SERVER: 127.0.0.1#53(127.0.0.1) (UDP)
 |    ;; WHEN: Sun Oct 25 01:40:31 UTC 2026
 |    ;; MSG SIZE  rcvd: 48
 |
 |> No no no no.  DNSSEC, i do not dig that myself.
 |> But yes, likely there the new algorithms will improve the
 |> situation a lot.
 |
 |For larger zone's like .COM there's work in progress to implement Merkle
 |Ladder DNSSEC, with just the tree roots signed with ML-DSA-44.

Oh.  See i have no idea.  This is interesting, frightening, and
shocking.  I would say stop exploding complexity, wait for the new
algorithms, they come soon enough.
(Sure i am talking too loud for my level of expertise in that
area, still it would be the right choice.)

 |> (In times where certificate rotation intervals get shrunken all
 |> the time, coming over with PQ for DNSSEC in 2026 seems very
 |> strange, sorry.)
 |
 |Not everything one does not yet understand is "strange" or a "mistake".

Poul-Henning Kamp has a nice email signature for decades:

    Never attribute to malice what can adequately be explained by
    incompetence.

I always loved it.
Does that hit back now, huh!  Better not think about that.

 |> But likely before 2030 i would at least think, if they get chosen
 |> in 2027.
 |
 |You need to include multi-year latencies for LTS releases of the
 |libraries to appear in LTS releases of an OS, which are then adopted
 |by enterprise users and operated for up to a decade.

That is a good point.
*However*, really, as long as the algorithm fits under the OpenSSL
interface they use, they will need to upgrade on algorithm
breakage, will they.  Yes, they will.
Not an argument for going MLDSA44 in my opinion.  Now i looked
when FIPS 2024 standardized that: August 2024.  That is
a thrilling two years.  Without any intent of sarcasm, if i add
two years onto 2027, i still fit 2030 pretty well!  And again, the
implementations of those already exist, in parts for years, ...
i think so.

 |> Some/most are already implemented in an OpenSSL compatible
 |> variant, in that PQ playground library where also the others came
 |> from (or where first).  I have read the communication on the FAEST
 |> algorithm (there).
 |
 |Implementation in liboqs has little relevance in real-world operations.

I have *zero* overview over that scene, whereas you are so deep
into this material, even as a top developer, .. right.  Kudos,
really...  Just one thing i can add here, our German secret
service plus, they (maybe, in parts: unfortunately) have contracts
with non-public law a.k.a. business, which write software, and
these are really involved there (for decades, i think the earliest
message i saw somewhere was from 2015), so i would *possibly* lift
your "real-world" to a certain "visibility level".  But that is
fine.  (Again, i am absolutely no cryptographer.)

 |> But one thing is plain: i will not add a DNS/UDP blowing algorithm
 |> in 2026 when 2027 brings algorithms which easily fit there,
 |> including ("normal") DNSSEC plus ("normal") DTLS, at least with
 |> EDNS.
 |
 |Nobody is suggesting that *you* or any other "typical user" switch
 |to MLDSA44(18).  It is in a pilot phase to explore the deployment
 |challenges.

See i hate that.  It has devastating effects to software
complexity and leads to quite lengthy and, sorry, pretty low
quality discussions in the second, third and further down
technical levels, like this DKIM WG.  And the WG's DKIM2 as
a technical monstrosity (not having read -07, saw R. Clayton
announcing -08 even), it really has other corners that should be
fixed.

It is just like that super-dirty Battery car monstrosity.  It
resulted and results in massive and very dirty and expensive
infrastructural changes, looting of resources in very fragile, or
even before untouched environments, and since the end of the 80s
or at latest beginning of the 90s every expert knows that hydrogen
and fuel cells it will be.  But so it goes.  What damage!

Ciao!!

--steffen
|
|Der Kragenbaer,                The moon bear,
|der holt sich munter           he cheerfully and one by one
|einen nach dem anderen runter  wa.ks himself off
|(By Robert Gernhardt)
_______________________________________________
Ietf-dkim mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to