Re: [v6ops] Vicious circle [ULA precedence [Thoughts about wider operational input]]

[Nick Buraglio <[e-mail]>]

I concur, Kevin. This has been the main driver for me to push that this
issue to be taken up. The work myself and others have been tasked with
encompasses the entire stack from application to optical transport, and it
is absolutely the applications and end devices that have the most work to
be done for IPv6 and IPv6-only, orders of magnitude more than the
networking, easily. Coming from primarily a service provider and high
performance networking background, I knew it would be the case but had no
idea just how extensive the gaps are until the work began about 18 months
ago.

nb



ᐧ

On Tue, May 3, 2022 at 8:45 AM Kevin Myers <[e-mail]>
wrote:

> I agree with both Gert/Michael would add that we need a focus on IPv6
> *systems* and not just IPv6 networks. Applications, compute and storage
> have not largely not been brought into an IPv6 world even in service
> providers.
>
> Enterprises focus on applications to support business goals (and thus
> indirectly the network) whereas in ISPs, the network is the business. This
> is why there is such a fundamental gap between what enterprise and ISPs
> care about - IPv6 is largely a solved problem in networking that just needs
> implementation, but it's not really a solved problem at all in system
> design. As Gert mentioned, a lot of work is needed to build better IPv6
> experiences in software.
>
> If we are going to support both worlds (which is the only real path
> forward for the success of IPv6) then IPv6 systems must become of paramount
> importance and not just a hyper focus on networks. It's all important.
>
> -----Original Message-----
> From: v6ops <[e-mail]> On Behalf Of Michael Richardson
> Sent: Tuesday, May 3, 2022 8:15 AM
> To: Gert Doering <[e-mail]>
> Cc: [e-mail]; 6man list <[e-mail]>
> Subject: Re: [v6ops] Vicious circle [ULA precedence [Thoughts about wider
> operational input]]
>
>
> Gert Doering <[e-mail]> wrote:
>     > On Tue, May 03, 2022 at 08:53:43AM +0200, Philip Homburg wrote:
>     >> So the conclusing seems to be: we have no clue how much enterprise
> is
>     >> spending on telecommunications, but because enterprise has a large
>     >> market cap, it must be important.
>
>     > My take is something else.  I want to relieve software developers
>     > (like, me) from having to bother with dual-stack, and focus on good
>     > IPv6 support.
>
> I would go further and suggest that there are efforts required to make
> certain communications models work in IPv4, which are much easier in IPv6.
>
> The ubiquity of IPv6-LL for instance makes many discovery and local
> connectivity problems significantly simpler.  It's a shame that docker went
> all IPv4 inside, because IPv6-LL is essentially ideal for all the
> container<->database connections that are brokered.
>
> I think that there are certain niches where IPv6 can be used even if there
> isn't Internet connectivity.
>
> --
> Michael Richardson <[e-mail]>   . o O ( IPv6 IøT consulting )
>            Sandelman Software Works Inc, Ottawa and Worldwide
> _______________________________________________
> v6ops mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/v6ops
>