On Mon, 1 Sep 2008, Ralf Weber wrote:
> Well we tested it as good as we could in our small lab
Ahh. This is where your engineers are supposed to consider theory, in
this case RFC1546 and RFC1812. Did you tell your senior management that
RFC1546 explicitly states that Anycast isn't suitable for TCP? Did you
inform your senior management that the Anycast TCP promoters are mostly
operations people (as opposed to credible engineers)? Did you tell them
that the Anycast promoters have financial conflicts of interest for
making their claims, and that they have no theoretical backing for their
assertions? Or did you leave all that out?
> and we did test tcp and we did have problems with it in one setup as
> described earlier. We did overcome this problems in the lab and now we
> are actively monitoring it (tcp connection to anycast ips). And we
> didn't got tickets of customers having problems with DNS TCP
> connection so far.
So you don't see the same source IP addresses repeated at multiple
anycast servers within a few minutes? You must have a very unique
setup.
> > Are your resolvers public?
> Of course not, I think there is a paper floating around here that this
> is a bad idea ;-).
Surprisingly, that paper is promoted by the same people promoting
Anycast DNS... There is no harm in public resolvers.
> > Would it be OK if I test them?
> Yeah will send you an separate mail about that.
Will do.
--Dean
--
Av8 Internet Prepared to pay a premium for better service?
www.av8.net faster, more reliable, better service
617 344 9000
_______________________________________________
DNSOP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dnsop
