L3: Google from DC via the Netherlands?

> Here's a theoretical solution to this problem that I'd like to open for
> discussion.
> In each location where a provider hosts their anycasted service, there
> is likely a local, non-anycasted IP address for each server.

There should be, yes.

> When
> receiving a DNS request that is not in the local cache, or has expired,
> make the new request on that local IP address interface, rather than on
> the anycasted IP address interface.

Yes. You probably have to do this in any case. Think about it. If you
have anycasted recursers in IAD, SJC, AMS, and HKG, and you're asking for
an answer hosted on a nameserver near IAD, and the query goes from the
anycast address to the near-IAD auth nameserver, then the response will
probably wind up at IAD, even if it was the HKG server asking. That will
not enable the HKG server to answer you.

You can probably hack your way around that issue by creative use of VPNs
and port assignments, but that's just a really poor-sounding solution.
Using the local IP address makes the right thing just magically happen.

> I'm curious to find out if others on the list know what other companies
> are using GSLB, and what the actual impact of anycasted DNS caching
> nameservers has on GSLB records. If enough people are using anycasted DNS
> resolution services, implementing a fix like this would reduce network
> traffic. By how much, I don't know.

The real problem is that if you're using an anycasted service, there is a
good chance that the recurser you're using is much further away from you
topologically than if you were just using a "local" recurser.

... JG