[dns-operations] Hijacking DNS traffic (Was: Re: new public DNS service:

Mark Milhollan mlm at pixelgate.net
Tue Nov 21 03:13:01 UTC 2017

Noel Butler wrote:

>So yes, there are non commercial reasons for doing that stuff. 

Rather you had effectively the same reason, and a commercial one: a good 
experience for your users.  Google needed people to get answers fast and 
without much faking (they take it further, yay!) so that other people 
can't replace their ads and services.  You needed people to get faster 
answers so they wouldn't call support (for "slowness"), and I bet you 
weren't against eliminating faked answers, at least those that would 
take your users to their old provider's search/help (and probably 
revenue) site.

Google uses their name and some FUD, some of which is certainly true, to 
convince people to use their service.  You used your power of position 
as the funnel though which the requests would flow.  Neither is very 
wonderful but each certainly seems defensible, yours not alone for "my 
network, my rules".

DNSSEC exists to prove the answers we get, who gives a flying fart where 
they came from so long as they validate.  Alas DNSSEC is far from 
universal, and some people do care or do at times.

But even DNSSEC presumes we run our own resolvers or have stub-resolver 

Running a local resolver used to be de rigueur but it is no longer done 
and mostly speed is the issue there, keeping folks using a larger 
entity's service -- memory and cpu are now usually non-issues.

There's not much security between the stub and a non-local resolver (or 
even a local one!), providing proof that the answers were/weren't 
spoofed (by your ISP or enterprise, much less malware) or where it is 
expected and condoned.


More information about the dns-operations mailing list