IPv8

Bender•about 2 hours ago

I keep seeing IPv8 (or maybe IPv4 V2) being suggested on HN. I do not see this happening in my lifetime. IPv6 while not perfect has taken 14+ years to reach almost 50% adoption most of which is from wireless and VPS providers and that has started to plateau. IPv8 if approved would have to work it's way through the daunting myriad of hardware obstacles, middle boxes, routers, operating systems and so much more that IPv6 had to complete. There are still a massive number of ISP's that need to implement IPv6 just to get people out of the CG-NAT mess. That and / or reclaim all the squatted, wasted and unused IPv4 space which is another topic in and of itself. I could clone the entire existing internet with the wasted space in IPv4 not even counting multicast. Just as an example with a couple orders from the head of the US and UK military several /8's could be freed up in a year. After that make each non ISP company that has anything more than a /20 justify each and every IP via on-site audits, quarterly. Part out ever "Future use" /8 and make non ISP companies justify their bids.

Given that IPv6 exists I can not imagine the people that integrated it now suddenly adding yet another routing solution. I would expect a majority of them to say something to the effect of, "Finish deploying IPv6 first as it already exists then we can talk about what gaps remain."

kstrauser•about 1 hour ago

> several /8's could be freed up in a year

But remember that each /8 freed with great time and expense (because organizations with a /8 can also afford a fleet of lawyers the size of Rhode Island) adds approximately 0.4% to the IPv4 pool. There aren't that many of those ripe for the plucking, either.

doener•about 2 hours ago

"Lucien,

I see no need for this IPv8. IPv6 was carefully engineered over many years and while not perfect, works and is deployed. What problem are you trying to solve? I seem to have missed that."

https://seclists.org/nanog/2026/May/9

mmooss•about 1 hour ago

That link is confusing. That message is from Gary Sparkes. Sparkes quotes a message from Jamie Thain [0], the author of the OP [1]. Thain quotes Lucien Hoydic who wrote that paragraph many messages prior [2] but for some reason Thain doesn't indicate it's a quote (e.g., Thain doesn't prefix the quote with ">") so it looks like Thain said it.

The best starting point is April 29th in the April archive, where Thain begins the discussion:

https://seclists.org/nanog/2026/Apr/date.html

The whole thread is people in the establishment mocking the idea, which is common to innovations good and bad, adopted and not.

[0] https://seclists.org/nanog/2026/May/6

[1] Per the FAQ: "Who is behind it?" "J. Thain (One Limited), author of draft-thain-ipv8-02 and the companion specifications covering routing protocols, RINE, ZoneServer, WHOIS8, NetLog8, and Update8."

[2] https://seclists.org/nanog/2026/Apr/130

lschueller•about 2 hours ago

Sorry, but how can an ip layer be an >integrated network management< ?

The github repo is completely empty. No draft / documentation? Looking forward to see a bit more of the technicalities

⚡ Community Insights

Discussion (3 Comments)Read Original on HackerNews