Hey, SiteFinder is back, again...

Mark Andrews Mark_Andrews at isc.org
Tue Nov 6 02:34:58 UTC 2007


> Mark,
> 
> On Nov 5, 2007, at 5:31 PM, Mark Andrews wrote:
> > 	All you have to do is move the validation to a machine you
> > 	control to detect this garbage.
> 
> You probably don't need to bother with DNSSEC validation to stop the  
> Verizon redirection.  All you need do is run a caching server.

	Yep.
 
> > 		dnssec-enable yes;
> > 		dnssec-validation yes;
> > 		forward only;
> > 		forwarders { <Verizon's caching servers>; };
> 
> Why bother forwarding?

	It was just to prove that you could detect this coming out
	of a ISP's servers.
 
> > 		dnssec-lookaside . trust-anchor <dlv registry>;
> 
> You forgot the bit where everybody you want to do a DNS lookup on  
> signs (and maintains) their zones and trusts and registers with <dlv  
> registry> (of which there is exactly one that I know of and that one  
> has 17 entries in it the last I looked).   You also didn't mention  
> that everyone doing this will reference the DLV registry on every non- 
> cached lookup.  Puts a _lot_ of trust (both security wise and  
> operationally) in <dlv registry>...

	There are also other lists of trust anchors.

	With 17 entries there arn't a lot of queries that need to
	be made to have the entire name space covered by cached
	NSEC records which DLV will use.
 
> > 	All lookups which Verizon has interfered with from signed zones
> > 	will fail.
> 
> Yeah, and Verizon customers would get a timeout (after how long?)  
> instead of a more quickly returned A (or maybe a AAAA) RR to a  
> Verizon controlled search engine.  Not really sure the cure is better  
> than the disease.

	But then you can log a complaint that DNSSEC doesn't work
	using their caching resolvers.  Or this just gives you
	the heads up to find the web form to change the servers
	returned by DHCP.  There is contributed code to do this
	linkage for BIND.  Or to manually update the forwarders.

	i.e. it's useful for those who use ISP's that havn't yet
	gone over to the dark side. :-)

> Also not sure what the point is -- most common  
> typos are already squatted upon and validly registered to a adsense  
> pay-per-click web page, typically a search engine (e.g.,  
> www.baknofamerica.com).  Seems to me the slimeballs have won yet  
> again...

	That's a different issue on a different battle front.
 
	Mark

> Regards,
> -drc
-- 
Mark Andrews, ISC
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: +61 2 9871 4742                 INTERNET: Mark_Andrews at isc.org



More information about the NANOG mailing list