6to4 status (again)

[Brian E Carpenter]

On 01/03/2013 07:56, Ignatios Souvatzis wrote:
> On Thu, Feb 28, 2013 at 07:15:12AM -0500, Brzozowski, John Jason wrote:
>> Keith we feel this way as such we deployed 6to4 relays.  Since our
>> customers were using the same we wanted to make sure it worked as well as
>> it could even though 6to4 offers a sub-optimal experience similar to other
>> tunneling technologies we ran trials for but did not deploy.  One of the
>> main differences with 6to4 is it is there whether we want it to be or not,
>> this is a bi-product of decisions that were made collectively year ago.
>>
>> We do feel that 6to4 traffic will go up before it goes down as native IPv6
>> deployment increases.  Only when native IPv6 reaches reaches a tipping
>> point do I believe 6to4 we will see a decline in 6to4 traffic and usage.
> 
> I wonder if I should try to add 6to4 *outgoing* on our router here,
> so that at least the back path from our Web server wouldn't touch 
> some anycast 2002::/16 device out there.

Yes. That's what Geoff Huston recommended a long time ago, and it
is also recommended by RFC 6343.

Switching off 6to4 at the client end is advisable, and (as David Farmer's
message explained) operating relays correctly is also advisable.
Switching correctly configured relays off clearly hurts users.

    Brian