Re: [DNSOP] DNSOP Presentation "The Camel"

[Robert Edmonds <[e-mail]>]

Paul Vixie wrote:
> Joao Damas wrote:
> > Camels are indeed great animals and they can be loaded until
> > eventually one more insignificant straw breaks their back. I guess
> > that is were Bert thinks the DNS is at now and I don’t completely
> > disagree
> 
> i was pretty horrified even before ECS. dnssec sentinels feels like friday
> the 13th part 37. there were danger signs saying "don't go here, there's no
> road, nothing to see, and no way back" and we ignored them.

Speaking of being horrified, compare the growth rate on the graph on
slide 2 in this presentation:

https://indico.dns-oarc.net/event/28/session/11/contribution/55/material/slides/1.pdf

With the growth rate on the graph in this article for the same time
period:

https://www.recode.net/2017/4/27/15413870/comcast-broadband-internet-pay-tv-subscribers-q1-2017

A hybrid resolution model that allows leaf queries for parts of the DNS
to be resolved direct-to-authority at endpoints would be very helpful
for some use cases (especially, CDNs) and could stall the insatiable
growth in resolution capacity needed at centralized resolvers. Not to
mention, such a model would provide a solid deprecation path for ECS.

I also note most but not all of the stuff Bert is talking about in his
slide deck are on the inter-server side of the protocol (resolver to
authority). But there are also other DNS camels that have been slowly
gestating in the browser world, e.g.:

https://plus.google.com/+WilliamChanPanda/posts/FKot8mghkok

https://bugzilla.mozilla.org/show_bug.cgi?id=1434852

-- 
Robert Edmonds