Ben Schwartz <[email protected]> writes:

> I am opposed to a "cache only" specification of LocalRoot.

So out of curiosity, would you actually like to state that
implementations shouldn't (SHOULD NOT) do that?

The localroot draft tries to be ambivalent about implementation, as
different implementations and deployments have different engineering
trade-offs to make.  Whether it is only cache-prefilling, or
cache-but-mark-never-expire, or build an entirely different authoritative
view to point at in parallel, etc, there are likely reasons for all of
them.

I worked very hard to make the localroot draft specify behavior
expected, not implementation mandates specifically.  I may not have
achieved perfection in that regard (and certainly, the WG, if we adopt
something may want to pick a way, or at least document "the likely best way")

-- 
Wes Hardaker
Google
_______________________________________________
DNSOP mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to