Re: [mif] [dnsext] 2nd Last Call for MIF DNS server selection document

[<[e-mail]>]

Now that more people are involved in discussions, this bare name / DNS
search list area seems to look like quite a deep swamp without clear IETF
consensus.

Perhaps we should discuss first if this particular topic (=section 4.6) is
needed in this document at all, because, after all, the focus is on
selecting the DNS server based on matching domains. 

Best regards,

	Teemu

> -----Original Message-----
> From: ext Keith Moore [mailto:[e-mail]]
> Sent: 21. lokakuuta 2011 17:10
> To: Savolainen Teemu (Nokia-CTO/Tampere)
> Cc: [e-mail]; [e-mail]; [e-mail];
> [e-mail]; [e-mail]; [e-mail];
> [e-mail]; [e-mail]
> Subject: Re: [mif] [dnsext] 2nd Last Call for MIF DNS server selection
> document
> 
> 
> On Oct 21, 2011, at 3:15 AM, <[e-mail]> wrote:
> 
> > Brian,
> >
> > Would the following text be then ok? Please note I changed the domain
> addition from SHOULD to MAY, if there is going to be attempt to
> deprecate/redefine/update search list logics. Or do you think it should
> remain SHOULD?
> > --
> > 4.6.  Interactions with DNS search lists
> >
> >   A node may be configured with DNS search list via DHCPv6
> >   OPTION_DOMAIN_LIST [RFC3646] or via DHCPv4 Domain Search Option
> >   [RFC3397].
> >
> >   A bare name (a name without any dots) MUST be first treated as a pre-
> >   DNS hostname or handled by other means that, as of this writing, are
> >   under discussion in the IETF and that are out of the scope of this
> >   document.  If the bare name resolution fails, the name MAY then be
> >   appended with the domain information.  If the bare name is appended
> >   with the domain information the described DNS server selection logic
> >   SHALL be utilized for the resulting name.
> 
> Associating MUST with undefined behavior makes no sense at all.
> 
> >   Resolution for the name containing any dots SHOULD first be attempted
> >   with DNS servers of all interfaces.  Only if the resolution fails the
> >   node MAY append the name with search list domain(s) and then again
> >   utilize improved DNS server selection algorithm to decide which DNS
> >   server(s) to contact.
> 
> Names containing dots SHOULD NOT (perhaps MUST NOT) be subject to
> searches.  They should already be considered fully qualified.
> 
> Just because a lookup "fails" does not mean that the name is not valid.
It
> could fail for temporary reasons, or because the TLD server wasn't
> reachable.
> 
> Back before there was a .CS TLD, searching on names containing dots was
> common.   Lots of computer science departments had .CS subdomains (e.g.
> cs.utk.edu used to be my mail domain), and people were accustomed to
> being able to send mail to moore@cs or [e-mail]).   Once the .CS TLD
> was defined it became obvious that domains containing any dots should not
> be subject to search.
> 
> Keith
>