<div dir="auto">Well yes, there must be a return path, otherwise Amos would see no traffic. Nobody is suggesting to *promote* 6to4, just no need to filter it.<br><br><div data-smartmail="gmail_signature">Regards<br> Brian<br> (via tiny screen & keyboard)</div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Wed, 15 May 2019, 21:56 Ole Troan, <<a href="mailto:otroan@employees.org">otroan@employees.org</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">> Anycast 6to4 needed to be assassinated, and that has more or less happened. If classical unicast 6to4 is still working for a few people, I don't really see any harm in it. Of course I agree that native is better.<br>
<br>
As far as I can tell there is really no flavour of 6to4 that can be deployed in a way where it works well.<br>
Even without the anycast IPv4 route, the far end IPv6 node would have to have a good route back to 2002::/16.<br>
If automatic tunnels are required 6RD contains all of the 6to4 issues within the ISP.<br>
<br>
Cheers,<br>
Ole<br>
<br>
</blockquote></div>