[gnso-rds-pdp-wg] For WG Review - Redlined Problem Statement

[Rob Golding]

On 2016-08-24 23:38, Stephanie Perrin wrote:
> I too would like to heed Chuck's advice, but have a burning question:
> Do registries have any way of distinguishing between bots and people,
> in their citing of stats?  If so, how?

Speaking as a Registrar, we can/do, so I'd certainly expect it of 
Registries:
* multiple queries in a very short space of time
* regular / consistent queries of the same domains
* queries of large numbers of domains from the same ip
etc

Most bots are usually pretty obvious -  there are the 'clever' and 
'responsible' bots/bot-authors/bot-using-orgs that spread load and not 
normally trying to "hide" who they are, so are given more latitude than 
others

Abusive levels of whois queries tend to get banned and even a handful of 
queries from a single source can get you blocked by many registrars and 
registries for a period

It can be a pain with com/net as the only current way to get the contact 
details (until the thick-whois project completes) as a gaining registrar 
is via the losing registrar whois server - some dont work reliably, some 
block you at very low levels from radar registered ips, lots dont work 
over ipv6 etc

Hit one recently who blocked you at *4 queries* in 10 minutes - a 
problem when the registrant was moving over 20 domains !

A lot of whois 'bots' aren't the right/best/appropriate way to get the 
data they're after anyway - a case of a little knowlege being dangerous 
- badly thought out drop-catcher type bots being one category that do 
excessive queries on the same domains which is not proper solution for 
what they're wanting to achieve

Rob