[AusNOG] IPv6: Where's my tunnel?

[Mark Smith]

>________________________________
> From: Mark Newton <newton at atdot.dotat.org>
>To: Noel Butler <noel.butler at ausics.net> 
>Cc: "ausnog at lists.ausnog.net" <ausnog at lists.ausnog.net> 
>Sent: Thursday, 7 March 2013 8:45 PM
>Subject: Re: [AusNOG] IPv6: Where's my tunnel?
> 
>
>
>
>I don't get it.
>
>
>Why are you people continuing to reward providers of IPv4-only transit by giving them money?
>
>
>Tunnels were a 2008 thing.  Unless you're talking about consumer broadband (which remains restrictive), there have been so many options for dual-stack transit for so long that there's very little excuse to not have at least ONE upstream link running it.
>

Consumer broadband is probably getting better if not already complete. Cisco have got it together, and it has been deployed in production by Internode. Around a year ago I saw Alcatel Lucent's presentation on their IPv6 broadband feature set and that was pretty complete. There is a presentation by Juniper towards the end of the month on this topic in Melbourne, and going by what is on the agenda, they probably have a pretty complete feature set too (and thinking about it XS4ALL in Europe used Juniper for their IPv6 broadband trial announced in around 2010). 

In broadband there is a bit of a mind set change needed, as now you're dealing with multiple layer 3 protocols, which don't need to be negotiated at the start of the session. For example, it is quite legitimate (and PPP supports this), for IPv6 to come up e.g., 5 hours after the PPP session came up, or for the PPP session to come up and no layer 3 protocols come up for a number hours, or for IP4 and/or IPv6 to be torn down a number of hours before the PPP session is torn down. In other words, PPP doesn't require the layer 3 protocols to be negotiated at the start of the PPP session or for them to be present at the end of it. Specific to IPv6, it is also possible for a DHCPv6-PD prefix delegation request to come many hours after IPv6 itself has come up. An example of when this could happen is when a host changes to being a router - this would happen when tethering is enabled on a smart phone (when IPv6 is supported over the 3G/4G connection, but that is a
 separate problem). The phone could switch back to being a host, letting the delegated prefix expire, all without the original PPP session being disrupted. The same model applies to Ethernet/"IPoE" implicitly.


I agree with Mark, vanilla, "non-authenticated", "non-quota" pay for pipe size, manually configured IPv6 services have been easily possible for easily 5 and more like 10+ years if the equipment supports IPv6, as it is just using basic IPv6 functionality to provide the service. Some of the reasonably recent IPv6 developments to support residential use could also make these a bit services easier to deliver e.g. prefix delegation via DHCPv6-PD, but they're really only relatively minor optimisations.

>
>Probably at least one of your upstreams actually does support it, but you haven't asked them to turn it on.
>

Yep, and that is a good way to create a self-fulfilling prophecy of IPv6 not being available. There is no demand for something if nobody demands it.

>
>Aren't you building IPv6 into your purchasing decisions anyway?
>
>
>   - mark
>
>
>
>
>_______________________________________________
>AusNOG mailing list
>AusNOG at lists.ausnog.net
>http://lists.ausnog.net/mailman/listinfo/ausnog
>
>
>