[v6ops] Re: Dynamic addresses

[Daryll Swer <[e-mail]>]

Gert

So basically, my understanding is, you've had zero personal exposure to,
and zero professional network consultancy projects that deals with the
impacts of *ever-changing* non-persisted ia_pd in an ISP network, across* N
*number of ISPs (remember, I don't work *for* ANY ISP, I work *with* MANY
ISPs globally, in multiple economies).

You have a strongly biased view on this subject in my personal and
professional opinion.

There's nothing but cons and problems with prefixes that changes often (per
5 hours, 6 hours, 24 hours or upon a quick OLT reboot, a quick fibre
splicing outside the user's home, a short packet loss in misconfigured MPLS
LSPs, that was quickly fixed by the ISP, but now the prefix changed and
there goes my kill in XYZ game).

Also, I don't understand why “SSH” is the “only” example of “application
impacted by dynamic prefixes”, has nobody advocating for dynamic prefixes
here, used VoIP? Nobody plays multiplayer games? And nobody actually worked
with multiple ISPs on this issue AND spoke to end-users' negative
experiences on the ever-changing prefix?

To those who advocate dynamic prefixes with the idea that “Other than SSH.
The average Joe won't notice anything”, please, do your homework instead of
assumptions, one small example from a quick Google search:
https://www.reddit.com/r/ipv6/comments/108iz77/dynamic_ipv6_from_isp_breaks_my_ipv6_connection/

(This is part of the larger discussion about IPv6 PI, of course, and
> why it does not make sense to give every barbershop wifi a PI /48,
> visible in global routing - because it does not work, and is not
> *needed* either)
>

I never saw a good reason for “/48 for everyone” approach. But I see no
problem with ISP static /56, where static = n number of years as a minimum
(10 years+), not seconds, hours, days, weeks, or months.

*--*
Best Regards
Daryll Swer
Website: daryllswer.com
<https://mailtrack.io/l/bd5ca3686315764199afc97a1144d48a674505de?url=https%3A%2F%2Fwww.daryllswer.com&u=2153471&signature=97c5dc22f6bd102c>


On Thu, 15 Aug 2024 at 02:28, Gert Doering <[e-mail]> wrote:

> Hi,
>
> On Thu, Aug 15, 2024 at 06:19:58AM +1000, George Michaelson wrote:
> > Surely before arguing for constant renumbering you should explain
> > remediation for the downsides of constant renumbering?
> [..]
> > Normally I find I agree easily with what you say. I think you are hand
> > waving consequences on your own experience with stable addresses as a
> norm.
> > If you have active solutions and live behind a dynamically addressed sub
> > prefix of some other delegate it would be useful to share how you make
> > inbound ssh work reliably.
>
> I do have a static v6 network at home, which has been static for
> roughly 23 years - so I see the consequences of assuming "addresses
> are static", like "addresses are being put into config files all
> over the world".  Should I ever cease working for this very ISP, it
> will take me months to get my network back into a working shape...
>
> Thus, not assuming stable addresses, and ensuring software copes better
> with change, and more automatic, will cause less pain in the long run.
>
> Now, of course I ssh home - from a few bastion hosts, because firewalling,
> and I do not expose my home network to "the world".  Making the DNS
> involved more dynamic would certainly be doable, but I am lazy.
>
> But, what you and I do, with SSH, is really of no relevance for the
> Internet at large.  99.99% of the end users out there do not know
> what SSH is, or why they would want a stable address - they use devices
> that seamlessly roam between wifi and LTE networks, keeping "connections"
> active, and all that - whether it's MPTCP underneath, some sort of
> VPN that can handle roaming endpoints, or just very short HTTP sessions,
> why would they care?  "Spotify keeps streaming when I leave home",
> this is a relevant metric.
>
> (This is part of the larger discussion about IPv6 PI, of course, and
> why it does not make sense to give every barbershop wifi a PI /48,
> visible in global routing - because it does not work, and is not
> *needed* either)
>
>
> ... I do have customer networks that are connected to other ISPs, with
> dynamic and semi-static /56s, and I use an Android phone as LTE hotspot
> with always-changing /64 on the wifi, so I *do* see where it hurts,
> and where it doesn't.   My Mac Laptop assigning itself a new privacy
> address when it roams between "living room" and "office space" wifi
> APs - because they have different SSIDs and it assumes "different network,
> then" - *this* is breaking my SSH sessions many times a day.  Which is
> annoying, because I have to press the hotkey for "please restore my SSH
> sessions to the standard set of hosts and reattach screens there"...
>
> Gert Doering
>         -- NetMaster
> --
> have you enabled IPv6 on something today...?
>
> SpaceNet AG                      Vorstand: Sebastian v. Bomhard, Ingo
> Lalla,
>                                            Karin Schuler, Sebastian Cler
> Joseph-Dollinger-Bogen 14        Aufsichtsratsvors.: A. Grundner-Culemann
> D-80807 Muenchen                 HRB: 136055 (AG Muenchen)
> Tel: [phone]         USt-IdNr.: DE813185279
>