>This message starts a DNSOP WG Call for Adoption of: >- draft-wkumari-dnsop-localroot-bcp >*OR* >- draft-hoffman-rootcache
In my opinion, both drafts should not be adopted. I have arguments why draft-hoffman-rootcache should not be adopted, but I'll leave them out until requested. I think draft-wkumari-dnsop-localroot-bcp solves a problem that does not really need solving. We have RFC 8806, an informational RFC which has roughly two parts, the first part explains local root and the second describes how to implement local root. The first part is fine. It is good to have a reference to the concept of local root. For the second part, resolvers tend to implement local root in a different way. So that part is not as useful. But this is fine, it is an informational RFC. Resolvers that implement local root differently do not violate an standards track RFCs. So do we need to do something about this? First, I think local root is not a BCP. Some operators use local root, which is fine. Many don't. And that is fine as well. There is (in my opinion) at this time no need for a specification that recommends local root. Does a standards track specification really help? What we have today is that resolvers re-use existing mechanisms for local zones to implement local root. A standards track specification is more likely to get in the way. Finally, Section 3.1 (Identifying locations from where IANA root zone data can be obtained) of the draft. The current root hints are deeply embedded in DNS. Today, every valid zone, including the root zone, has to have an apex RRset. This RRset can be resolved to addresses. The root hinsts can be constructed from information in the root zone itself. Section 3.1 changes that in a way that is mostly undefined. This means that to implement this draft in the real world additional policy is needed for authors of software and for operators on how to stay current on the list of sources of the root zone. This draft requires a lot of effort to solve a problem that in the operator world doesn't seem to exist. _______________________________________________ DNSOP mailing list -- [email protected] To unsubscribe send an email to [email protected]
