George Michaelson wrote:
I am less sure the UDP/TCP thing reduces to "no" I see no reason any more to assume session cost is too high for a globally deployed DNS. I suspect what DNSOPS and a hypothetical directorate thinks about DNS is less impactful (sorry, hate that word) than what embeds in Android devices. ...
i don't think there's any decision support coming from those observations. the internet is still older and larger than the web, and i expect this condition to continue, but if i'm wrong, it will still never be small enough to ignore, or indeed, small in any way.
we should support android, and the web. that won't change what the rest of the internet needs to keep doing, for reasons of its own, for at least the lifetimes of the children of all members of this mailing list.
let me illustrate. a shell user on a virtual container somewhere, in 2049 let's say, who needs to pull down an update in single user mode so she can get out of single user mode, will need to do a dns lookup. yes, we have proven that we can use persistent/stateful initiators in that use case, and that we can parse json. but can that become mandatory, because the web came, and was large, and didn't need datagram transport?
i say no. the internet predates the web and might outlive it. -- P Vixie _______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop
