[prev in list] [next in list] [prev in thread] [next in thread] 

List:       ipng
Subject:    Re: [IPv6] [EXTERNAL] Re: Developing a solution to a problem that shouldn't exist (Re: ULA vs. 1918)
From:       "Manfredi (US), Albert E" <albert.e.manfredi () boeing ! com>
Date:       2023-06-18 22:16:21
Message-ID: 1fcb9e0076be437da56613b9b091045d () boeing ! com
[Download RAW message or body]

-----Original Message-----
From: ipv6 <ipv6-bounces@ietf.org> On Behalf Of Brian E Carpenter

> If two ULA prefixes are in the same network, after a network merge, ULA->ULA \
> *should* be preferred. In that case, either a high precedence for fd00::/8 or high \
> precedence for both ULA /48s would work.

If you're careful to block ULAs at the edges of your own network, yes. I didn't get \
Ole's point at all, that ULA should be preferred, *unless* we're certain that ULAs \
are not sneaking their way into the Internet. There's hardly a guarantee that ULAs \
are unique over the Internet. We don't even know for sure what kind of PRNG people \
might use, to create the 48-bit random prefix. Or even if they follow that rule.

When Ole points out:

> ULA is going to be treated as globally reachable. I.e. follow default route.

I'd say, we have an unsafe situation, in the real world, which can't be ignored? I'm \
asking, because I don't understand these points.

Many have said that ULAs are finding their way into the Internet, much as RFC 1918 \
have done. I suppose one can conclude, lessons learned from IPv4 suggest you prefer \
GUA always, for that reason. Unless you are administering your own network and you \
are diligent in blocking ULAs at the edges of your network. That's a big caveat \
though, isn't it?

Bert

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


[prev in list] [next in list] [prev in thread] [next in thread]