So all rules are dynamic (where can they go wrong) and you not only require all hosts to be In DNS but you also require LLDP to be turned on? You are overcomplicating everything. Shane
On May 3, 2026, at 6:05 PM, Jamie Thain via NANOG <nanog@lists.nanog.org> wrote:
From host a to host b in the ipv8 world host b ipv8 address is
0-192.168.2.3 and in a network 65535 is on a different segment.
Local segments cannot share the same ip address.
Host b looks like the DNS address of host a Xlate8 makes a translate rule to the destination is the zone server, the source is the card and makes a new rule where the destination is the IPv8 interface of the zone server and the destination is the IPv8 192.168.2.3
192.168.2.3 send the IP address of the zone server which gets translated to the IP address of the IPv8 and it has the source address of the zone server and IPv8 so it sends packets back to that and then back to 192168.2.3
Now barring that every IPv8 network gives a DHCP name to every interface, and learns a c name by the LLDP name and if the device also wants to propagate some kind of name like the web server then you can put it in the LLDP name and it will capture it or you can configure a c name to that IP address.
Jamie
On Sun., May 3, 2026, 5:05 a.m. Owen DeLong, <owen@delong.com> wrote:
Host A is at address 65534:192.168.2.3 Host B is at address 192.168.2.3
How, exactly, do you think these two hosts magically exchange packets?
Host B has no IPv8 stack and no ability to produce an IPv8 packet with the extra 32 octets. Host A must be 100% backwards compatible to live up to your claims.
Yes, I realize my example involves a private ASN and a public address. Substitute overlapping public addresses if you prefer, the point remains.
Owen
On May 2, 2026, at 05:56, Jamie Thain via NANOG <nanog@lists.nanog.org> wrote:
Mikael
And what isn't backwards compatible if you know of something. I'm all ears.
Jamie
On Sat., May 2, 2026, 6:18 a.m. Mikael Abrahamsson, <swmike@swm.pp.se> wrote:
On Sat, 2 May 2026, Jamie Thain wrote:
Ipv10 Was A dual stack idea.
Everything that isn't IPv4 will be a new stack. This was realised already in the 90s, and that's why we have IPv6 that's now close to 30 years into its deployment. Adding another stack in the mix won't help.
If you would have done what you're doing back in mid 90s, it might have helped. You're 3 decades too late.
Again since you can't do the analysis steps then just send me a tape recording.
There is nothing to analyze. You're proposing something that isn't backwards compatible, and you don't seem to understand this is what you're proposing. Same as IPv10 guy.
So start implementing, you might understand it then.
-- Mikael Abrahamsson email: swmike@swm.pp.se
_______________________________________________ NANOG mailing list
https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/Z4YOGRDC...
_______________________________________________ NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/UMJJIVVS...