ietf
[Top] [All Lists]

Re: [DNSOP] new DNS classes

2017-07-07 13:19:00
Mark,

On Jul 6, 2017, 11:56 PM -0700, Mark Andrews <marka(_at_)isc(_dot_)org>, wrote:
Or you could stop trying to reinforce the myth that new RR types
are hard to deploy. They really aren't.

Please stop trying to minimize the amount of work here. They really are. Not 
for you, but for the folks who make domain names available for use by 
non-technical types.

Then change domain hosting providers and tell them why or run you
own master server or use a service which allows for dynamic updates
which shouldn't care about the record types. There are plenty of
DNS providers that will slave content.

The number of folks who even understand what you've written above, much less 
can actually do it, is insignificant in the context of seeing significant 
adoption of new RRs.

As a DNS server vendor

You need to look at the wider world.

The DNS server side of deploying a new RR is the easiest part and while 
necessary, it's no where even close to sufficient to getting a new RR deployed. 
It's like trying to move the automobile industry to hydrogen fuel: building a 
car that runs on hydrogen is easy, relatively speaking. However to actually do 
anything useful with that car, you need need to modify the fuel delivery 
infrastructure, which is vastly harder, involves a vastly larger set of 
players, very few of which have any sort of financial incentive to make the 
change. Standard chicken-or-egg problem.

ICANN funded John Levine to 
develop https://metacpan.org/release/JRLEVINE/Net-DNS-Extlang-0.1 to try to 
make it easier to break the chicken-or-egg problem by hopefully giving the 
folks who provide the UIs by which domain names are managed by non-geeks better 
tools. However, that's just the first step -- the domain name managers still 
need to be encouraged to use them.  Pretending this is an easy problem isn't 
helping.

Regards,
-drc
<Prev in Thread] Current Thread [Next in Thread>