Disclaimer
Mind: it's a draft... so it's "work in progress." yada yada yada...
These are my thoughts on this draft which at the time of writing this is 1 day old.
But IPv6...?
Didn't saw a success since 1998 when it was released as also declared in the draft:
IPv6 was developed to address exhaustion. After 25 years of standardisation and deployment effort IPv6 carries a minority of global internet traffic. [...]
IPv8 is intended to solve this by having IPv4 backwards compatibility.
Management Philosophy
The central operational concept is the Zone Server — a paired active/active platform running every service a network segment needs: DHCP8, DNS8, NTP8, NetLog8, OAuth8, WHOIS8, ACL8, and XLATE8.
When first reading this I was amazed but thinking about it further made me skeptical. Stuff like DNS, NTP, WHOIS, ACL and XLATE makes sense and OAuth is also nice for businesses but centralized log collection? That just makes me question security and stuff like guest networks: what is really logged with this? Because this draft is in its early stages this is not fully defined. Probably will be interesting.
Address Space
This draft is strange. The IPv8 Address Format is r.r.r.r.n.n.n.n where r.r.r.r is the ASN and n.n.n.n is the host address. This means there is space for 2^32 (4,294,967,296) ASNs with 2^32 (4,294,967,296) host addresses each. I think an r.r.r.n.n.n.n.n approach would be better because this we could have 2^24 (16,777,216) ASNs - which is more than enough - with 2^40 (1.09951163e12) host addresses each.
Lets do the math for an ISP with about 8 million (2^23) customers:
r.r.r.r.n.n.n.n
for an ISP with 2^23 connected households would be
2^32 / 2^23 = 512 hosts per household.
r.r.r.n.n.n.n.n
for an ISP with 2^23 connected households would be
2^40 / 2^23 = 131072 hosts per household.
It could just be that it is just because I like IPv6 but 512 hosts sounds like not-so-much.
Even/Odd Addressing — New in -01
Draft-01 introduces a proper spec for something that was only hinted at in -00: the even/odd Zone Server redundancy model. Every subnet has two Zone Servers at .254 (even) and .253 (odd). Even-addressed hosts route via .254, odd-addressed hosts via .253. Dual-NIC hosts get one even and one odd address from DHCP8 — one per NIC — giving them active use of both gateway paths simultaneously.
This is genuinely elegant. Redundancy and load distribution fall out of the addressing convention itself, with no stateful load balancer required. The A8 DNS record spec is updated accordingly: responses should be an even/odd pair so clients can open parallel streams across both paths.
XLATE8 gets the same treatment: when an IPv4 client connects to an IPv8 host, the XLATE8 gateway can distribute connections across the even and odd addresses of the destination. IPv4 clients get the benefit of IPv8 load distribution transparently.
WHOIS8 — Now Explicitly Critical Infrastructure
In -00, WHOIS8 was mentioned as a route validation mechanism. Draft-01 is more direct: WHOIS8 is described as a critical infrastructure service underpinning the entire global routing model. BGP8 route acceptance is conditioned on a valid WHOIS8 record. No valid record, no route installation.
This raises a question the draft doesn't fully answer yet: what is the failure model? If WHOIS8 is as central as the spec implies, its availability and governance become at least as important as the routing protocols themselves. Who runs it? What happens when it's unreachable? This feels like the most consequential open question in the suite right now.
CGNAT
The section "Requirements for a Viable Successor" states that CGNAT will continue to work. While this is essential for this draft's protocol success I just feel like it should also be included that CGNAT isn't required and NAT in general wouldn't be required with this. Just to prevent ISPs putting NAT in front of this like with IPv6.
Cost Factor (CF) Routing
One of the more interesting parts of this draft is the Cost Factor (CF) metric, combining the dynamic composite path quality of EIGRP, the accumulated cost model of OSPF, and proportional load balancing - in a single open versioned algorithm that operates end-to-end across AS boundaries.
The physics floor is a neat idea: no path can appear better than the speed of light over the great circle distance allows. Anomaly detection for free.
But I have questions. CF is measured from TCP session telemetry - what happens with encrypted traffic? And the economic policy component means money flows directly into routing decisions. Who controls that? When major carriers enter their own peering costs into the algorithm, this effectively results in commercial favoritism toward certain routes. This is structurally similar to the net neutrality debate, only buried deeper in the stack.
Conclusion
Sounds nice! The even/odd model in -01 is a real improvement - it's the kind of thing that looks obvious in hindsight. The WHOIS8 elevation to critical infrastructure is the right call architecturally, but it needs a governance and failure-mode story before this can be taken seriously as a deployable spec.
My three problems remain:
- Missing definition of NetLog8 (will certainly come soon).
- Address formatting (4 billion ASNs? In case half the world wants one? WHY?)
- Net neutrality is really important (especially for our democracy).
And I'd add a fourth: WHOIS8 governance - the spec can't stay silent on this much longer.