My initial observation is that this I-D leans quite heavily on RFC 9859, "Generalized DNS NOTIFY". It also leans somewhat less heavily on a few current internet drafts by the same authors (I-D.berra-dnsop-keystate, I-D.leon-dnsop-signaling-zone-owner-intent, and I-D.berra-dnsop-announce-scanner.) I assume these drafts are intended to travel together. I did not review those drafts as part of this review. The introduction expects that the reader is familiar with DNS NOTIFY, DDNS, and SIG(0). It also _mentions_ RFC 9859, but my reading of this I-D suggests that the reader also needs to be quite familiar with that RFC as well. Perhaps say, instead of saying "In addition this Internet-Draft borrows heavily from the thoughts and problem statement from [RFC9859]", which is an understatement, say something like "In addition the reader should also be familiar with [RFC9859], in particular with the definition and use of the DSYNC resource record defined in section 3." I say this because on my initial read of the draft, I was very confused about the first step in this protocol: the client queries for a DSYNC record under *some* QNAME. But what? It turns out that this is quite well defined in RFC 9859 section 3. Thus, this draft is confusing if you haven't read that RFC, or don't remember it. My next observation is that while section 4 says: > This document does not define a new update mechanism. The child's update to the delegation information is an ordinary DNS UPDATE [RFC2136], secured as a SIG(0)-signed transaction [RFC3007] [RFC2931]; the NS, glue and DS changes it carries, and the way the UPDATE Receiver processes them, are exactly as specified there. the actual mechanism that the UPDATE Receiver processes updates is actually unlike any actual Dynamic DNS implementation that I'm aware of. This is acknowledged in the I-D later in section 4: > Another concern is that traditionally DNS UPDATEs are sent to a primary nameserver, and if the update signature verifies the update is automatically applied to the DNS zone. This is often not an acceptable mechanism. The recipient may, for good reason, require additional policy checks and likely an audit trail. Finally, the change should in many cases not be applied to the running zone but rather to some sort of provisioning system or a database. So, I find the first statement a little jarring. I think the intent is to point out that the DNS UPDATE mechanism is using the exact same _protocol_ as previously defined, even though the UPDATE Receiver implementation is necessarily different. In section 5, we have: > Apart from defining a new scheme to specify the mechanism "UPDATE" (rather than the mechanism "NOTIFY") this document does not say anything about what Qname to look up or what RR type. The UPDATE mechanism should use exactly the same method of locating the target of the UPDATE as is used for generalized NOTIFY. I found this paragraph is confusing. I don't think the phrasing "this document does not say anything about what Qname to lookup" is even correct -- it *does* say, but not directly. Perhaps be more direct: >> This document defines a new DSYNC scheme specifying "UPDATE" (rather than "NOTIFY"), but otherwise follows the same method as generalized NOTIFY for determining the Qname to look up (e.g., {child}._dsync.{parent} and variants as described by Section 3 of [RFC9859] I have some other observations and nits, in no particular order; * I'm not convinced that any of section 3 is useful. I think the whole section could be removed. * In Section 7.1 paragraph 1, it says: > A single UPDATE used for delegation maintenance is not expected to also carry KEY changes. Why not just say that a single UPDATE MUST either contain delegation changes OR KEY changes? The protocol might work with them combined, but wouldn't this simplify implementations? Or conversely, either don't mention it, or explicitly allow the combination? * In section 9.1, it says > The child's SIG(0) public key is, in the common case, discoverable in DNS: either at the child apex, or at the _sig0key.{child}._signal.{nameserver}. name in the nameserver's zone, or via multi-query convergence against the child KEY RRset. What is "multi-query convergence"? I assume this is a concept from CDS bootstrapping or one of the other referenced I-Ds, but it seems out of the blue here. * In section 9.4.1.1, it says > If the child zone is DNSSEC-signed (including a signed delegation via a DS record), then the KEY record containing the SIG(0) public key located at the zone apex can be looked up and validated by the DNS UPDATE Receiver. It might be obvious to many of us, but what does "validated" mean? You might want to spell out that if the returned, DNSSEC-validated KEY RRset has at one KEY that matches the KEY in the bootstrapping UPDATE message, then it should be considered validated. Or maybe you mean something else? * In section 9.4.1.3 > Should a "registry" parent want to support this mechanism (as a service to its unsigned children) then a likely model is that the target of the DNS UPDATE is operated by the registrar (or possibly that the DNS UPDATE is forwarded to the registrar). The registrar performs its normal verifications of a change and then transforms the update into an EPP [RFC5730] transaction to communicate it to the registry. This seems generally applicable to the entire process, although the same can also be said of CDS/CSYNC. That is, the registrar could process the DNS UPDATE and convert it to an EPP update in a similar fashion to converting a CDS scan change into an EPP update. It isn't clear that this model works directly with DSYNC, though. Obviously, the registry could forward the DNS UPDATEs to various registrars, but would it be possible for clients to directly find their *registrar's* UPDATE processor? I don't see how. * In section 9.5.1, > at-ns: This is an indication that the UPDATE Receiver supports automatic bootstrap of the SIG(0) public key by publishing the KEY record both at the child zone apex AND at the special name _sig0key.{child}._signal.{nameserver}. Why does "at-ns" *also* mean "at-apex"? The example below this section shows both "at-ns" and "at-apex". * In section 9.6: > Once the parent (or registrar) UPDATE Receiver has the key, the child can update it via a DNS UPDATE just like updating the NS RRset, the DS RRset or the glue in the parent zone (assuming a suitable DNS UPDATE policy in the parent), i.e., only the initial bootstrapping of the key is an issue. What is the difference between the normal DNS UPDATE and the bootstrap? The only thing I can think of is which KEY is used to sign the UPDATE. In the bootstrapping case, it is the KEY in the UPDATE. In the "normal" DNS UPDATE, it must be the prior trusted KEY. I think it would be useful to point this out in this section. * A nit: in section 2 > Because of the drawbacks of CDS and CSYNC scanners they are unlikely to be able to fully automate the maintenance of delegation information in all parent zones. I would add a comma: >> Because of the drawbacks of CDS and CSYNC scanners, they are unlikely to be able to fully automate the maintenance of delegation information in all parent zones. Or this could be reworded >> The drawbacks of CDS and CSYNC scanners make them unlikely to be able to fully automate the maintenance of delegation information in all parent zones. * Another nit: in section 3.1 > Both NOTIFY and UPDATE are messages, in DNS packet format. I would remove the comma here: >> Both NOTIFY and UPDATE are messages in DNS packet format. * And another nit: in section 9.4.1.2 > *Operator coordination via HSYNCPARAM.* Why is this bold? Should it be a sub-section header? Or at least on its own line?