When RFC 6540 was written 12 years ago it was reasonable to expect that many services didn't yet support IPv6 and that the IETF lacked the purchasing power to convince vendors to make changes. However in 2024 we've had a decade of major governments and corporations including IPv6 support requirements in RFPs. We still have another decade or more of many vendors having a long tail of features still catching up with IPv6 support, but lacking IPv6 should be considered a bug at this point. We're at the leading edge where it's reasonable to build certain classes of applications as IPv6-only (sometimes needing some IPv4 transition mechanism, but wanting to keep use of that to a minimum).
In that context, it seems reasonable for the IETF to have a reference to RFC 6540 in all RFPs and to have supporting both IPv6 and IPv4 as a MUST.
Having an exception process may be necessary but hopefully rare, should be tied to no other options being available, and should be tied to a concrete plan and roadmap from the vendor for addressing.
Github is one clear exception we've had for awhile, and this is painful to many people in the community as it is a roadblock to IPv6-only services. Indications are that progress is being made there to dual-stack, and it would be nice to avoid further regressions.
But either way, I agree this is an area where the IETF needs explicit clarity and policy rather than just periodic mailing list discussions.
Erik
