DNS Reliability

Mon Sep 16 20:45:56 UTC 2013

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 13/09/13 12:45, Valdis.Kletnieks at vt.edu wrote:
> On Thu, 12 Sep 2013 14:03:44 -0600, Phil Fagan said:
>> Everything else remaining equal...is there a standard or
>> expectation for DNS reliability?
>> 
>> 98% 99% 99.5% 99.9% 99.99% 99.999%
>> 
>> Measured in queries completed vs. queries lost.
>> 
>> Whats the consensus?
> 
> Remember to factor in Duane Wessel's work that showed that
> something like 98% of the DNS traffic at the root servers was
> totally bogus?
> 
> Maybe you need to factor in "broken queries not answered, and
> offenders slapped around with a large trout"?  Because if it's
> busted requests you're sending towards the root, they're going to
> count against your completed/lost ratio in a really bad way.
> 
> Anybody know if people have cleaned up their collective acts since
> Duane did that paper?
> 

Wearing a different hat, I had the chance to rerun that analysis with
data from 2008 (original paper is from 2003) and the number were still
around 98%

http://www.caida.org/publications/presentations/2008/wide_castro_root_servers/wide_castro_root_servers.pdf

Cheers,
- -- 
Sebastian Castro
DNS Specialist
.nz Registry Services (New Zealand Domain Name Registry Limited)
desk: +64 4 495 2337
mobile: +64 21 400535
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with undefined - http://www.enigmail.net/

iEYEARECAAYFAlI3bfYACgkQWyqRrHcQWTkagwCeOaShzFH1i8q9Y34/cybV6bUY
qBYAn1A8JPgNJqH6mijUFN7+4ufybJqZ
=X7UE
-----END PGP SIGNATURE-----