multiple prefixes

Phil Mayers p.mayers at
Mon Feb 11 16:47:38 CET 2013

On 11/02/13 15:42, Tim Chown wrote:
> On 11 Feb 2013, at 14:52, Phil Mayers <p.mayers at IMPERIAL.AC.UK>
> wrote:
>> Personally I think caring about client choice of address in
>> "normal" LAN networks is not time well-spent, but it's something
>> reasonable people can disagree on.
> Basically it's either 'we'll control all the addresses in use' or
> 'we'll accept devices can pick any address and we'll adapt our
> management and monitoring to work with that'.

Isn't the former actually "control all addresses in use and accept that 
clients without DHCPv6 are forever lost to us"?

> Given the proliferation of devices on our network, and that the vast
> majority are user-owned, we've taken the latter approach.  Whether
> university campuses are "normal" is another question :)

Good point; based on what I've spent the day doing, I suspect "normal" 
is something we're very far from...

On a more general note, we've always managed clients IPv4 address/DNS 
entries quite closely; we've not done that with IPv6 and noticed no ill 
effects, and given the growing domination of wireless/BYOD, I wonder if 
those days aren't coming to a close entirely.

More information about the ipv6-ops mailing list