[NANOG] Re: IPv6 Legacy IP Warning Stickers

["Colin Stanners (lists) via NANOG" <nanog () lists nanog org>]

As an example, MikroTik RouterOS has, just 2 months ago, finally added IPv6
FastTrack (a certain type of hardware offloading) in their beta software.
Until now, FastTrack was IPv4 only, so for much of their routers' usage
cases, the maximum IPv4 throughput would be 3-4x faster than IPv6. That
massive performance difference still applies to the shipping software of
units being purchased today.

https://www.reddit.com/r/mikrotik/comments/1i6kwpe/ipv6_fasttrack_support_ad
ded_in_v718beta_testing/

(Unfortunately that is only in RouterOS7 which has just recently become
stable enough for important uses; for the last few years, providers have
been staying on the old 6.x branch, and away from the latest hardware that
is RouterOS7-only, for stability).


-----Original Message-----
From: Brandon Martin via NANOG <nanog () lists nanog org> 
Sent: March 26, 2025 3:26 PM
To: nanog () lists nanog org
Cc: Brandon Martin <lists.nanog () monmotha net>
Subject: [NANOG] Re: IPv6 Legacy IP Warning Stickers

On 3/26/25 14:00, Andrew Latham via NANOG wrote:
> Today yet another ISP is running Fiber in the utility easement at the 
> street. I checked and they do not offer IPv6 or have ANY IPv6 peering.
>
> I have offered the hard to find IPv6 Legacy Warning stickers on my 
> Redbubble profile. About a month before any meeting or event I see a 
> bulk order for Amish IPv6 stickers designed by Phil Benchoff and hard 
> to find post Google+. I have the markup/profit set to the lowest 
> setting and have made maybe $8 over 5+ years.
>
> Linky:
> https://www.redbubble.com/i/sticker/Legacy-IP-Warning-by-gringomalvado
> /38585698.EJUG5
>
> Ranting a bit as I have four ISPs boxes at the street and two of them 
> do not have IPv6 in 2025. Converting my frustration into finding Phil 
> and sending him some coffee money.

Until consumers care, bulk-subscriber driven residential ISPs probably won't
care.  They have to provide IPv4 anyway, and most of the startups are stuck
running CGNAT to do it already, so IPv6 is just another operational hassle
for them.

I don't like it, but it's the way it is.

The major content providers could do a fair bit on this front.  They could
start by supporting IPv6 at all (sadly many don't), and for those that do
support IPv6 (which is a lot of them - and thank you!), they could nudge
their customers into getting it enabled on the grounds that it may improve
their experience (which is true since it can and usually will bypass
overloaded CGNATs).

OTOH, I just had an issue where I ended up disabling IPv6 for a customer and
improving their all-important speed test metrics by 50% across the board to
all test providers who support IPv6.  My peering is pretty symmetric, but I
assume their (older but far from obsolete) router has hardware offload for
IPv4 but not IPv6 since that seems to be the bottleneck in all cases.
Obviously I could tell them to get a new router, but we know how that's
going to end.

--
Brandon Martin
_______________________________________________
NANOG mailing list
https://lists.nanog.org/archives/list/nanog () lists nanog org/message/SAKPWS7V
VM7LOC54NNMNNW6PO6ZLYFDU/

_______________________________________________
NANOG mailing list 
https://lists.nanog.org/archives/list/nanog () lists nanog org/message/Z2KTYNMILNHQ2WH6OKPE7D4W3NUOJH2H/