|
On 9/29/2017 6:48 PM, Brian E Carpenter
wrote:
A BCP with the right tone and focus might be useful.First of all, I agree with those who have said this should be a BCP, if published. BCPs are the way we publish IETF process rules. Secondly, I think many of the comments about the tone and slant are correct. What we want to stop is work on solutions that are *specific* to IPv4, and to chase down and elminate any cases where successful IPv6 operation depends on the presence of IPv4. I disagree. We need to consider IPv4 work as "maintenance mode", which can easily include solo IPv4 adjustments and/or include IPv4 support in new protocols that also support IPv6. Neither necessarily need involve transition or deprecation. "no new work" or "no IPv4-specific work" both assume that IPv6 is a superset of IPv4, which it is not. We're still wrangling with aspects of IPv6 that actually are evolving back into IPv4-like approaches, e.g., limits to the length of the header chain and problems supporting fragment traversal of routers. Joe |