Re: [ga] Verisign rushes to roll out typo-squatting system on Monday

[Karl Auerbach <karl@xxxxxxxxxxxx>]

On Mon, 15 Sep 2003, Dan Steinberg wrote:

> It seems to me that if there is a value for TTL that is undefined and 
> set by different parties at will without even a necessary expectation 
> that a TTL be a particular value for that party...then the act of 
> calling it a 'technical standard' is deceptive, no?

The TTL on a DNS record is set in the zone file, which means that it is
set by whoever maintains the zone file.  I'm not sure in the
registry/registrar situation whether it is the registrar or the registry
that is in control of the value that is applied.

As for the value of the TTLs - There is I believe, an RFC of suggested
practices regarding things like TTL values, but it is not, I believe, an
RFC that is an Internet Standard (not all RFCs are standards, just as not 
all standards used on the net are in RFCs.)

And there are operational practices that aren't well documented - like 
turning down the TTL values before an anticipated change in a record, thus 
reducing the period of confusion caused by any residual information that 
remains floating around after the change.

One measurement that we missed with the .org switchover of servers was an
actual measure of the cross-fade behaviour when TLD servers are
decommissioned and new ones activated.  Louis Touton and I discussed
setting this up, but we both got distracted.

		--karl--