[ipv6hackers] "Stick to limited IPv6 deployments, businesses warned"

[Fernando Gont]

On 09/08/2012 02:49 PM, Gert Doering wrote:
> On Thu, Sep 06, 2012 at 05:01:26AM +0000, Jim Small wrote:
>> SLAAC is a done deal.  The question is, will network and OS vendors 
>> extend their SLAAC implementations to support RDNSS.  From what I can 
>> see, it seems like the answer is no.  RDNSS is cool, it's nice for labs 
>> and other setup types, and it's a published standard.  However, it's 
>> pretty much worthless if routers and mainstream O/S (e.g. Microsoft/Apple) 
>> don't support it.  
> 
> JFTR, iOS (Apple) supports RDNSS since iOS 4.<something>, so quite a while.
> 
> But besides that, you're right - Cisco is not implementing it "because
> we have DHCPv6 and all the clients do (stateless) DHCPv6 as well", and
> it's hard to find reasonable counterarguments.

The counter argument probably being that rather than having two
alternate autoconf mechanisms, we have two mandatory autoconf
mechanisms, since non of the can provide all the encessary information
needed for autoconf (e.g., one doesn't provide EDNSS, while the other
does not provide default routes).

Unfortunately, religious wars about SLAAC vs. DHCPv6 didn't help us at all.

Cheers,
-- 
Fernando Gont
SI6 Networks
e-mail: fgont at si6networks.com
PGP Fingerprint: 6666 31C6 D484 63B2 8FB1 E3C4 AE25 0D55 1D4E 7492