Re: draft-ietf-v6ops-addr-select-ps-02

[Arifumi Matsumoto <arifumi@nttv6.net>]

Hi,
thank you for your comment.

Iljitsch van Beijnum wrote:
> I'm reading draft-ietf-v6ops-addr-select-ps-02 and I found something 
> that I find somewhat troublesome:
>
> 2.2.1. IPv4 or IPv6 prioritization The default policy table gives IPv6 
> addresses higher precedence than IPv4 addresses. There seem to be many 
> cases, however, where network administrators want to control the address 
> selection policy of end- hosts the other way around.
>
> I have a hard time believing this is true. Why would anyone want to 
> enable IPv6 and then leave it unused? And if it is true, it's highly 
> problematic: if IPv6 capability is available but not used, this means 
> that problems (that are common when deploying new technologies) will 
> remain unseen and therefore unfixed.

We've found some operational reasons for IPv6-IPv4 priority control.

- As mentioned some time before at this ML, smooth and conservative
 IPv4 to IPv6 transition might be one reason.

 The transition looks like:
  IPv4 only network,
   -> IPv4 and IPv6 network with IPv4 prioritized.
    -> IPv4 and IPv6 network with IPv6 prioritized.
     -> IPv6 only network

- Another case might be like you have temporariliy extremely poor IPv6
 access line and want to use that only when connecting to IPv6-only
 host.

However, I agree that a new technology, such as IPv6, should be
generally prioritized than the old one.