To:
"D. J. Bernstein" <djb@cr.yp.to>
cc:
dnsop@cafax.se
From:
Pekka Savola <pekkas@netcore.fi>
Date:
Fri, 18 Jul 2003 10:43:14 +0300 (EEST)
In-Reply-To:
<20030717183131.45666.qmail@cr.yp.to>
Sender:
owner-dnsop@cafax.se
Subject:
Re: regarding the respsize draft: preferring glue of certain types
On 17 Jul 2003, D. J. Bernstein wrote: > (1) The resolver can't reach IPv6 addresses. This means that the > initial addresses have to be A---otherwise they _do not work_. > Similarly, all the subsequent DNS server addresses used by that > resolver have to be A---otherwise they _do not work_. You're assuming that if you reach one IPv4 address, you can reach all. > (2) The resolver can reach IPv6 addresses. In this case, if a DNS > server supports both IPv4 and IPv6, why would the resolver be > interested in the DNS server's IPv4 address? What's the benefit? You're assuming that if you reach one IPv6 address, you can reach all (, and that you have no use for IPv4 addresses for fallback). Also, assume the resolver is caching data. Someone else with different transport protocol capabilities asks the resolver about the addresses of these servers -- and get a partial response. Building assumptions like these doesn't sound like the way to ensure the robustness of DNS. -- Pekka Savola "You each name yourselves king, yet the Netcore Oy kingdom bleeds." Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings #---------------------------------------------------------------------- # To unsubscribe, send a message to <dnsop-request@cafax.se>.