[DNSOP] Re: New draft on collision free key tags in DNSSEC

[Edward Lewis <[e-mail]>]

On Jul 25, 2024, at 20:34, Shumon Huque <[e-mail]> wrote:
>     https://datatracker.ietf.org/doc/html/draft-huque-dnsop-keytags-00 <https://datatracker.ietf.org/doc/html/draft-huque-dnsop-keytags-00>

I suggest changing this:

Colliding key tags impose additional work on a validating resolver, which then has to check signatures for each of the candidate set of keys identified by the Key Tag.

to:

Colliding key tags may be used to hide an attempt by a zone manager to maliciously lure a validator into an overly computationally and futile attempt to validate data, a form of denial of service.

I think “that” is the problem, not just the fact that there may be multiple keys with the same tag.  Operationally, in benign situations, I’ve seen two keys “collide” twice - not often and a shallow problem.  Only once had I noticed it while it was live - the first time it came and went without anyone complaining, with evidence that the operator did notice and fixed it.

As far as the document beyond that, I’d suggest “enforcement” be placed at the primary server - the one place the zone is assembled and assigned a serial number.  A zone could be refused loading if tags collided, pushing back on the admin/operator whose new key probably caused the collision.

OTOH, this is a DNSSEC key management tool issue, not a resolver/server issue.  Yes, resolver (validator) is impacted but cannot fix the problem.  With multi-signer in mind, there are multiple, discreet (not coordinated) key generators in use, the only common point (or first common point in the flow) is at the primary DNS server.

I don’t see how this can be DNS security algorithm specific, how collision avoidance could be made part of a “DSA”’s definition.

Ed