What I see is a hack that would be ineffective in a system that actually
drove serious aggregation. If we really drove prefixes to
aggregatability, then only those few exceptions which were highly
unusual could use the equivalent of IPv4 Anycast.
(Note, this is different from IPv6 anycast which, as I understand, is
more like naming a portion of the map, rather than trying to name an
arbitrary set of scattered end-nodes.)
So yes, I think anycast as captured by IPv4 treatment (like anycast DNS)
is an abnormality, not a useful direction.
If we want to make that service a first class citizen, we will have to
do so very differently, and recognize the costs as with multicast.
Yours,
Joel
William Herrin wrote:
On Wed, Apr 8, 2009 at 10:56 PM, Joel M. Halpern <[email protected]> wrote:
The last analysis I saw is that anycast is in some sense a fiction. That is,
it is not meaningful to describe an anycast IP "address" as an address
according to any of the useful definitions of address.
Rather, it is a name for a collection of things.
We have merely overloaded the routing system so as to route on that
particular kind of name.
Joel,
So you see an abnormality here, not a pattern in the evolution of routing?
Regards,
Bill Herrin
_______________________________________________
rrg mailing list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg
