RE: IPv6 DNS Autoconfiguration

[Pekka Savola <pekkas@netcore.fi>]

On Mon, 14 Jul 2003 matthew.ford@bt.com wrote:
> > I want a spec which is simple and clear, and less than 15-20 
> > pages long.
> 
> That would certainly be nice, but I don't think we should base our decision
> on the length of the specifications for the proposed solutions ;-)
> 
> Benefits of the RA approach I can see are:
> 
>  o no need to implement DHCP clients in all hosts
> 	- this could be important for very simple network devices

   o no need to run DHCP clients in all scenarios
        - I certainly would not have any need to run DHCPv6(lite) on IETF 
WLAN, discovering the recursive DNS is just enough.

If you roam a *lot* e.g. with a PDA and use Mobile IPv6, you may not 
desire all that many functions from your care-of networks.

>  o no need to solicit data explicitly
> 	- hosts will either hear an RA with the relevant data, or send an RS
> and get the name server details as part of the RA response. I don't support
> the idea of an 'active' mode whereby hosts can explictly request the DNS
> data. If the router(s) is/are configured with the name server data and are
> configured to advertise it then they will do so. Explicit requests only make
> sense if we anticipate explicit requests for other types of data, and I
> think this should be absolutely out-of-scope.

I'm not sure I'm 100% sure what's best myself, but there is one advantage 
to the "active" model (it seems): it may be easier to handle if your RA/RS 
processing happends in the kernel, but processing DNS info (including 
rewriting /etc/resolv.conf) would have to be in the userland.

-- 
Pekka Savola                 "You each name yourselves king, yet the
Netcore Oy                    kingdom bleeds."
Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings

#----------------------------------------------------------------------
# To unsubscribe, send a message to <dnsop-request@cafax.se>.