This is idiotic. I'm sorry/not-sorry for the bluntness, but it just is. It's "100% backward compatible", but one of the first projects is:

> Kernel build of the IPv8 stack

> - Linux kernel branch implementing the IPv8 packet header (RFC §6)

> - AF_INET8 / sockaddr_in8 socket API surface

> - ARP8 dual-probe neighbour capability discovery

> - 8to4 tunnelling for IPv8-over-IPv4 transit

So aside from having to reinvent everything, including kernel-level packet header processing, it's A-OK! It also adds:

> Security by protocol

> - East–west isolation via ACL8. North–south egress validated against DNS8 + WHOIS8.

...for those who were tired of being able to ping other hosts on your LAN without, let's see, checking with a layer 7 protocol first.

In an interview, the creator said[0]:

> Another IPv8 feature is what Thain calls a “Zone server” that his draft explains “runs every service a network segment requires: address assignment (DHCP8), name resolution (DNS8), time synchronisation (NTP8), telemetry collection (NetLog8), authentication caching (OAuth8), route validation (WHOIS8 resolver), access control enforcement (ACL8), and IPv4/IPv8 translation (XLATE8).”

Boy, you thought people had opinions on systemd? Wait until they get a peek at this all-your-eggs-in-one-basket critical infrastructure.

I mean this seriously: there's no part of this that I want anything to do with. Zero. Nothing. I'm only being so harsh here because I don't want to leave any wiggle room of doubt that perhaps I think any of it is redeemable, because it ain't.

[0]https://www.theregister.com/networks/2026/05/12/veteran-netw...

I feel like this completely misses the entire point (or at least half the functionality) of ASNs.

ASNs give you an identity at the peering level that can be used to dynamically provision routes; so it is used to evolve (over time) where a given network-level address lives. Making it part of the address now makes that mapping vastly less flexible.

The list of things it breaks is breathtaking:

- Transferring a subnet between providers can't be done (the ASN part changes).

- Providing the same services on multiple providers doesn't work (different ASNs can't advertise the same address).

- Various types of multicast domain management break (global networks that use IBGP for internet multicast routing need to figure out their prefixes).

Then there's the bootstrapping problem with the mix of layer 3 / layer 7. And JWT used everywhere? And no integrated solutions for things like DNSSEC (warts and all)? No apparent attention to distributed implementations of (now consolidated) core infrastructure?

Say what you want about systemd, at least there is arguably someone applying a different (but consistent) design to a real problem. This "proposal" feels like someone with limited understanding of the design principles behind the Internet opened up ChatGPT and asked it to fix "problems" that they barely understand.

I try not to be mean to the well-meaning, but... well... some people aspire to operate at a level for which they are entirely unqualified.