Re: [mif] draft-ietf-mif-current-practices-00

> -----Original Message-----
> From: Dan Wing [mailto:dwing@cisco.com]
> Sent: Tuesday, April 06, 2010 11:52 AM
> To: Dave Thaler; 'Hui Deng'; Gabriel Montenegro
> Cc: mif@ietf.org
> Subject: RE: [mif] draft-ietf-mif-current-practices-00
> 
> 
> 
> > -----Original Message-----
> > From: Dave Thaler [mailto:dthaler@microsoft.com]
> > Sent: Tuesday, April 06, 2010 10:06 AM
> > To: Hui Deng; Dan Wing; Gabriel Montenegro
> > Cc: mif@ietf.org
> > Subject: RE: [mif] draft-ietf-mif-current-practices-00
> >
> > Hui is correct, Windows has per-interface DNS server lists
> configured.
> >
> > It then uses a host-wide "effective" server list for an actual query,
> > where the effective server list may be different for different names.
> >
> > On Windows the per-interface suffix is actually termed the
> > "connection-specific DNS suffix" to distinguish it from the
> > "primary DNS suffix" of the machine.  I think that's why
> > "interface-specific" was repeated in the first bullet.
> 
> 
> 
> In draft-montenegro-mif-multihoming, there are two purposes and terms
> that
> seem to be intermingled using the term "DNS suffix".
> 
> One purpose is the suffix for non-FQDN names, like "payroll" or "mail",
> which will have a suffix added to them (e.g., example.com).  

That's what windows calls the "DNS Suffix Search List" (see the
sample output I sent previously below).  It's called the
"domain search list" in other places (like RFC 3397), or just
"search list" (RFC 1123).

> The
> other purpose is deciding which DNS server will be be sent a query for
> a certain FQDN (e.g., queries for *.example.net go to one DNS server
> and queries for *.example.com go to a different DNS server).

Another purpose is deciding which DNS server will receive a dynamic 
update for a name with a certain suffix (e.g., Windows supports dynamic
updates for the primary DNS name, and optionally also the connection-
specific DNS name of the machine).

> 
> 
> In draft-ietf-mif-current-practices-00, which is the WG document
> that seems to have boiled down draft-montenegro-mif-multihoming,
> but draft-ietf-mif-current-practices-00 also does not clearly
> separate the two purposes.

Yep

-Dave
> 
> -d
> 
> 
> > Example on Windows, extracted from "ipconfig /all" output:
> >
> > Windows IP Configuration
> >
> >    Host Name . . . . . . . . . . . . : dthaler-win7
> >    Primary Dns Suffix  . . . . . . . : ntdev.corp.microsoft.com
> >    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> >    Node Type . . . . . . . . . . . . : Hybrid
> >    IP Routing Enabled. . . . . . . . : No
> >    WINS Proxy Enabled. . . . . . . . : No
> >    DNS Suffix Search List. . . . . . : ntdev.corp.microsoft.com
> >                                        redmond.corp.microsoft.com
> >                                        ntdev.microsoft.com
> >                                        dns.corp.microsoft.com
> >    System Quarantine State . . . . . : Not Restricted
> >
> > Wireless LAN adapter Wireless Network Connection:
> >
> >    Connection-specific DNS Suffix  . : hsd1.wa.comcast.net.
> >    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> >    Description . . . . . . . . . . . : Intel(R) Wireless WiFi
> > Link 4965AGN
> >    Physical Address. . . . . . . . . : 00-1D-E0-34-4F-6F
> >    DHCP Enabled. . . . . . . . . . . : Yes
> >    Autoconfiguration Enabled . . . . : Yes
> >    Link-local IPv6 Address . . . . . :
> > fe80::4853:4753:9d8d:3b45%13(Preferred)
> >    IPv4 Address. . . . . . . . . . . : 192.168.0.195(Preferred)
> >    Subnet Mask . . . . . . . . . . . : 255.255.255.0
> >    Lease Obtained. . . . . . . . . . : Monday, April 05, 2010
> > 10:19:02 PM
> >    Lease Expires . . . . . . . . . . : Tuesday, April 06,
> > 2010 10:19:02 PM
> >    Default Gateway . . . . . . . . . : 192.168.0.1
> >    DHCP Server . . . . . . . . . . . : 192.168.0.1
> >    DHCPv6 IAID . . . . . . . . . . . : 335551968
> >    DHCPv6 Client DUID. . . . . . . . :
> > 00-01-00-01-12-0C-E2-7A-00-1E-37-CC-8D-DD
> >
> >    DNS Servers . . . . . . . . . . . : 2001:df8:0:1::25
> >                                        192.168.0.1
> >    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> >    NetBIOS over Tcpip. . . . . . . . : Enabled
> >
> > -Dave
> >
> > > -----Original Message-----
> > > From: Hui Deng [mailto:denghui02@gmail.com]
> > > Sent: Monday, April 05, 2010 7:40 PM
> > > To: Dan Wing; Gabriel Montenegro; Dave Thaler
> > > Cc: mif@ietf.org
> > > Subject: Re: [mif] draft-ietf-mif-current-practices-00
> > >
> > > DNS server always has specific interface related information,
> > > but the final DNS server will still be host based, I wouldn't say
> it
> > > is not correct.
> > >
> > > one example would be you have internet connection and vpn
> connection
> > > at the same time,
> > > good VPN implementation will always rely on VPN DNS server
> > information
> > > for Internet connection.
> > >
> > > -Hui
> > >
> > > 2010/3/31 Dan Wing <dwing@cisco.com>:
> > > > Section 3.2.1.3 of describes the DNS configuration of Windows,
> and
> > > says:
> > > >
> > > >  "Interface specific DNS configuration can be input via static
> > > >   configuration or via DHCP.  It includes:
> > > >
> > > >   o  An interface-specific suffix list.
> > > >
> > > >   o  A list of DNS server IP addresses."
> > > >
> > > > It is curious that the first bullet repeats "interface
> > specific", but
> > > the
> > > > second bullet does not repeat it.  A reasonable interpretation is
> > > that the
> > > > second bullet is not interface-specific, but the lead-in sentence
> > > says this is
> > > > interface-specific.  I was hoping
> > draft-montenegro-mif-multihoming-00
> > > would
> > > > clarify, but it doesn't.
> > > >
> > > > -d
> > > >
> > > > _______________________________________________
> > > > mif mailing list
> > > > mif@ietf.org
> > > > https://www.ietf.org/mailman/listinfo/mif
> > > >
> >
> 