[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: I-D ACTION:draft-ietf-ipv6-unique-local-addr-08.txt



Stephen Sprunk wrote:
...
also imo - this whole idea is a clear and present danger to the Internet
(assuming that IPv6 gets general deployment)


I disagree. The risk of these non-aggregatable prefixes appearing
in the default-free BGP4 table in exchange for lots of money is the same
as the risk of *any* longer prefix from PA space doing so. The danger
exists as long as rich enterprises are willing to pay for routeability.
There are other things we are doing (renumbering procedures, multi6, the
NAP draft) to try and deflect this danger, but ULAs don't increase it.


I'll note that ARIN is, at this very moment, debating whether to allow PI assignments directly to end sites; if that proposal is approved, they will pay only $100/yr for a prefix that any ISP can be reasonably expected to accept. The costs, in both time and effort, of getting every relevant ISP in the world to accept PA or ULA routes will be significantly higher.

Not quite that simple. Firstly those "PI" prefxes would in fact be taken out of PA space, because that's all the registries have today. Secondly they would only be for people already allocated an AS number, and the registries estimate that about 10k prefixes would be the maximum in practice. However, you're correct that this would provide a much more attractive safety valve than inappropriate use of ULAs.

   Brian

--------------------------------------------------------------------
IETF IPv6 working group mailing list
ipv6@xxxxxxxx
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------