Why used DHCPv6 when RA has RDNSS and DNSSL?

[Ola Thoresen]

On 4/1/20 7:16 PM, Gert Doering wrote:

> Hi,
>
> On Wed, Apr 01, 2020 at 05:53:02PM +0200, Bjørn Mork wrote:
>> Lorenzo Colitti <lorenzo at google.com> writes:
>>
>>> I'm not sure that the folks asking for IA_NA would be happy with IA_PD
>>> though.
>> Why don't you just try and see?  You have nothing to lose AFAICT.
> I've said it on IETF discussions and will happily say it again - this
> is a particular corner-case that might benefit developers running stacks
> of VMs in VMs on their laptops, but I fail to see a real world argument
> why this would be beneficial.


I really dont see this as "corner cases".  Maybe a few years ago, but in 
todays world, more and more apps are distributed as contianers and my 
guess is that we will see even more stuff like that in the future.

IF we could easily get prefixes assigned to "workstations" (or whatever 
we call them) I can also imagine that it will quickly be used by various 
kinds of sensors, "IoT"-devices etc. to build "personal area networks".  
Think stuff like your wireless headset, local file sharing between 
devices, personal ip-cameras and so on.  Lots of devices that either use 
USB or Bluetoth or other similar connections today could easily connect 
to one of your local PD-prefixes and benefit from a simpler and more 
robust connection over IP.

I also fail to see why the cost of implementing PD in a network would be 
significantly higher than doing all the ND snoping and logging you would 
need to track individual addresses - _especially_ in an enterprise network.


/Ola (T)