Re: IPv8 / BGP8 / CF

[Dan Mahoney via NANOG <nanog () lists nanog org>]

> On Apr 29, 2026, at 12:03 PM, Jamie Thain via NANOG <nanog () lists nanog org> wrote:
>
> Hi All,
>
> My name is Jamie Thain I'm the creator of IPv8. It's not a hoax.

Back in the day, I came up with the idea that, as part of the IMAP specification, you could move a message to a special 
folder that would cause messages to be acted on server-side to create dynamic procmail/sieve/imapfilter rules.

Lots of places already use this for spam reporting, but I had greater ambitions, including blocking a specific sender, 
or to put a user in "timeout" for a while, or even trigger autoresponders.  One of the most useful features I 
envisioned was the ability to automatically mute a given thread before it took up more mental energy than I had in a 
given week.

Today, current-me is really annoyed at past-me for not fully implementing that last one.

===

Let me tell you a fun story of IPv6.  For pretty much *decades* while IPv6 existed, the only way to get v6 support on a 
router that you were already running with tight memory constraints, and often bought without TAC, was to run a beta 
version of the OS.  Cisco IOS was built for the network stack people are using today, not the one that people might use 
later.  There are lots of devices (literally every one of them) on the internet that have support for IPV4, and 
probably most of those will now support ipv6, but at the overhead of "twice the routing table size, twice the RIB, 
twice the peering to manage, twice the customer complaints if there's a weird routing issue".  Happy Eyeballs came WAY 
later than ipv6 did.

With IPV8, "now you have three problems".

Today, lots of vendors sell kit on the secondary market for which there are no firmware updates, and even if you wanted 
to buy a support contract to get that firmware, they won't sell to you.  To support your proposal, every one of them, 
all of them, would have to be upgraded.

The same reason people haven't broadly supported ipv6 is the same exact reason this proposal is a non-starter.  Perhaps 
in a green-field overlay internet (see also, the 6bone, the mbone, tor, LISP, AMT) you could use this, but it would 
still have to run on routers and switches that had no idea what it was, at the ip protocol layer.  You're never going 
to get that.

My day job deploys dual-stack, everywhere, in production, and we've been doing it for decades.  We think that's 
important to do.  We've definitely found some sites and corners of the world where one protocol acted *very 
differently* from the other.

We also find that 90 percent of the traffic we receive on both those stacks is "people who couldn't even figure out how 
to configure ONE protocol correctly".

-Dan
(Opinions are my own and do not reflect dayjob's opinion at all -- I am an independent network operator with my own 
ASN/IP space).
_______________________________________________
NANOG mailing list 
https://lists.nanog.org/archives/list/nanog () lists nanog org/message/FT3KDQCRBKK7HIN2A3BTN6WZ2X5UNOWE/