Re: [v6ops] The need for local-ipv4 socket transition solutions -- NAT64/DNS64 remains insufficient

[Ca By <[e-mail]>]

On Wed, Apr 22, 2015 at 6:59 AM, Ted Lemon <[e-mail]> wrote:

> On Apr 22, 2015, at 8:02 AM, Heatley, Nick <[e-mail]> wrote:
> > If everyone complaining here, would work towards getting rid of them
> rather than discussing who’s problem it is to put a workaround in place,
> we’d probably be half-way done.  We have people crawling the web, people
> having users and test networks, people who seem to be able to write enough
> emails on the topic that they could even direct it to the right people.
>
> The problem is that the people who are using IPv4 literals are not
> _here_.   If fixing this were up to the people who are _here_, it would
> already have been fixed.   Nobody _here_ thinks that using IPv4 literals is
> okay (I suppose I will be corrected if I am wrong).
>
> So while what you are saying would definitely be right if those people
> were here, since they are not, it isn't practical.
>
>

The key is that:

1) We can only "fix" what is in our control -- as a network operator
464XLAT is in my  toolbox -- thank you IETF / Android / Microsoft.

2) Any solution that involves creating a negative customer experience (xyz
website or abc app does not work), is a non-starter.  Please stop
suggesting pain on paying customers, no network operator will do it. No
appstore will do it (with teeth, in the near term)

3) As Ted noted, the "ipv4-only deployers" are here and they are not
moving.  They include Google (hangouts), Apple (facetime), Spotify,
Netflix, so on and so forth. They understand the issue, and they are not
running to fix it.  So what is an ipv4-poor network to do?  Well, let's
check the tool box:

1) CGN -- tried true, but really just kicks the can down the road -- oh,
and RFC1918 ran out soo....  i hear there is some IPv4 space that is not
announce on the internet i can use .... So, this is a pretty anti-social
option, but it works.

2)  464XLAT / MAP / DS-lite -- Yay!  i can give users unique IPv6 public
addresses, grow the real IPv6 internet, and let ipv4 die on the vine (x
decade process, but real meaningful ipv6 progress now)

3.  Buy IPv4 (does not scale well, also anti-social)

Keep in mind, the tool box only includes these 3 things.  Wishing some
vigilante titan network is going to rain fire down on IPv4-only apps is not
realistic at all.  Same is true for  "mega-mobile-phone eco-system", they
will also not bar IPv4-only apps in a timescale that is helpful.  And, they
too have customers that want to go to http://192.0.1.1 on the public
internet ... and no, they are not going enroll networks to switch to
IPv6-only without IPv4-socket support to reach 99% of internet while other
networks are delivering 100% of the internet.

Regards,

CB




> _______________________________________________
> v6ops mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/v6ops
>