[dns-operations] IPv6 PTR best practice

John Levine johnl at taugh.com
Wed May 9 16:27:10 UTC 2018

In article <Pine.LNX.4.64.1805082245080.19081 at byq-nqzva.cvkrytngr.arg> you write:
>> The *MASSIVE* number of /unneeded/ records would be astonishing.  Not to
>> mention the amount of resources that said records would consume.

Well, let's see.  My home LAN uses one /64 out of the /56 that my ISP
assigns.  Let's assume that each PTR record uses a modest 30 bytes
since both the name and the data can use label compression.  So 2**64
records times 30 bytes comes to about 500 million terabytes.  Seems
bulky, don't think it would fit in my garage.  And, of course, each of
my neighbors will need another 500M terabytes in their garages, too.

>Or synthesizing them on-the-fly, which doesn't need so many resources 
>but then we're back to whether the textual result is useful and the 
>arguments around that.

Some people in dnsop were floating a proposal for a BULK RRTYPE to
synthesize records on the fly.  Among its less awful flaws is that it
requires online zone signing.  It's not going anywhere.


More information about the dns-operations mailing list