[dns-operations] requirements for TLD servers

[Jim Reid]

On 22 Mar 2010, at 14:30, Peter Koch wrote:

> Agree it could, but disagree for the "should", unless you'd want to
> start an infinite endeavour.  Think "jurisdiction" or "constituency".

I can't see how these things can (or could) be linked in any way to  
what is a technical matter concerning engineering/operations. An IETF  
or OARC document is not legally binding on anyone and any  
"jurisdiction" or "constituency" is free to ignore it.

If some ccTLD(?) registry did ignore such an agreed engineering BCP  
and then their government/regulator not unreasonably started asking  
questions about RFC2870-bis compliance, that would be a National  
Matter. It wouldn't hurt too IMO if ICANN expected better DNS  
resilience from gTLDs. The current requirements are rather weak: just  
pass the CNNP test. I accept that some ccTLDs don't have the resources  
to match the best infrastructure. Even so, an agreed set of minimum  
requirements would give them something to aim at as well as something  
which could be referenced by RFPs, hosting contracts, etc.

Peter you seem to be saying that there shouldn't be something which  
documents a reasonable set of minimum requirements for important DNS  
zones because this might interfere with layer-9 silliness in certain  
"jurisdictions" or "constituencies". I hope I'm wrong about that.