[dns-operations] "Authoritative Name Server" at Wikipedia

Michael Hoskins (michoski) michoski at cisco.com
Wed Aug 8 22:48:21 UTC 2012


-----Original Message-----

From: bert hubert <bert.hubert at netherlabs.nl>
Date: Wednesday, August 8, 2012 3:23 PM
To: Mike Hoskins <michoski at cisco.com>
Cc: Jim Reid <jim at rfc1035.com>, David Conrad <drc at virtualized.org>,
"dns-operations at lists.dns-oarc.net" <dns-operations at lists.dns-oarc.net>
Subject: Re: [dns-operations] "Authoritative Name Server" at Wikipedia

>On Wed, Aug 08, 2012 at 10:08:52PM +0000, Michael Hoskins (michoski)
>wrote:
>>While we're tongue in cheek, everyone knows DJB invented "DNS done
>>right".
>
>In a major sense he did. You can still run djbdns from 10 years ago and
>not
>get hacked.
>
>Nobody else managed that.


In all fairness, it's a lot easier to avoid security flaws when you offer
a limited feature set that's frozen in time.  (The fact you think you can
run djbdns from 10 years ago seems to support the generally frozen state.)

I wouldn't run djbdns from 10 years ago though, even with far less
features than BIND Dan's had to pay up on his security guarantee.  :-)
Smart guy no doubt, but no one's perfect.  Thinking you can deploy
anything without careful monitoring and maintenance just makes one lazy.

That said, having extensively used both products (as well as others like
Unbound and PowerDNS) -- to each their own.  We're using some of the
"whiz-bang" features BIND provides, and it better integrates with another
product we're using.

Honestly, I got somewhat soured on DJB w/ qmail in an ISP environment -- I
was one of the first to hop on the postfix bandwagon and never looked
back.  I digress, that's not DNS.

-----Original Message-----
From: RijilV <rijilv at riji.lv>
Date: Wednesday, August 8, 2012 3:29 PM
To: Mike Hoskins <michoski at cisco.com>
Cc: Jim Reid <jim at rfc1035.com>, David Conrad <drc at virtualized.org>,
"dns-operations at lists.dns-oarc.net" <dns-operations at lists.dns-oarc.net>
Subject: Re: [dns-operations] "Authoritative Name Server" at Wikipedia

>On 8 August 2012 15:08, Michael Hoskins (michoski) <michoski at cisco.com>
>wrote:
>>PPS: Best practice suggestion -- don't name products after well
>>understood
>>industry terms.  :-(
>
>You mean like windows :P  that seems to be working out pretty well for
>them...


Sure.  I'd wager ANS worked out fine for Nominum too, but that wasn't my
point -- for the community / industry, avoiding a source of confusion
works out best.  In the end, that also works out best for the vendor.  M$
and Nominum aren't the only guilty parties here, it's all too common to
see the same terms overloaded again and again.  It's partially marketing,
partially laziness, and partially the fact we already have so many terms.
Regardless of the reason, it presents an unnecessary barrier to learning
for new folks.  When possible, I like to think it should be avoided.




More information about the dns-operations mailing list