Re: DoS and anycast

[Patrik F�ltstr�m]

On m�ndag, nov 4, 2002, at 15:52 Europe/Stockholm, Randy Bush wrote:

to be specific, i see no reason that the root cache ip address for
server X should not be anycast internally within multiple isps as
long as
  o the isps' anycast servers serve consistent data

Ack

  o the scope of the anycast routing announcement is bounded

Ack

  o if X is controlled by some entity, e.g. the iana, that the
    entity coordinates the mechanism(s) of zone distribution and
    authentication, and maybe even has means of testing content

I don't know if we say the same thing. Let me try to explain what I think:

If organization X is responsible for IP address Y today, i.e. X:Y is in the whois database, then X is also the organization people will call when wrong data comes from address Y, regardless of what copy of Y the wrong data comes from.

Because of this, for every Y, there must be an X which know where all copies of Y is, so they can see/detect which one have wrong data -- and fix it. The organization X should even have a warning system which warns them when one of the copies of Y give wrong data / is out of sync.

paf


#----------------------------------------------------------------------
# To unsubscribe, send a message to <[EMAIL PROTECTED]>.