[dns-operations] Quick analysis of TLD SOA's

Jake Zack jake.zack at cira.ca
Wed Jul 13 15:03:37 UTC 2011

I queried all TLD's for their SOA record and analyzed expire, refresh,  
and retry times.

Found a few things that seemed odd to me, but may well be perfectly  
normal and acceptable to the rest of the community, so I just wanted  
to throw it out there and perhaps hear others thoughts.

Checked 303 TLD's.
Average Expire Time in days: 22.37

BO: 18144007 - 210.00 days
TRAVEL: 15724800 - 182 days
EG: 12096000 - 140 days
AERO: 6048000 - 70 days
LA: 6048000 - 70 days
HM: 6048000 - 70 days
MO: 432000 - 5 days
UY: 432000 - 5 days
VI: 216000 - 2.5 days
BF: 86400 - 1 days
KM: 86400 - 1 days
AW: 60480 - 0.7 days
PW: 10800 - 0.125 days

...I understand large expire times (.CA's was set to 1 week, but as of  
next zone publish (11:00am EDT) will become 40 days due to this  
analysis)...but I'm wondering why some TLD's have opted for incredibly  
short (<1day) expire times.  Any thoughts on this?

Also, in comparing refresh time with retry time, I noticed that all  
but 3 TLD's have their refresh time set higher than their retry time.   
The 3 who break with this convention are:

ET: refresh(600) < retry(1800)
FM: refresh(1800) < retry(3600)
LA: refresh(900) < retry(1800)

In this case, though, I can't really think of a situation where one  
would intentionally do this.  Thoughts?

-Jacob Zack
DNS Administrator - CIRA (.CA TLD)

More information about the dns-operations mailing list