RE: RE: [dhcwg] New draft on DHCPv6 Relay Information Option and RADIUS Attributes sub-option

Great, I have no objection to your suggested name other than it is rather
long - perhaps it will become known as OPTION_RRA.

Are you or Ralph going to discuss this at the upcoming DHC WG session and
request it become a WG work item?

BTW, we don't need a DHCPv6 version of
draft-ietf-dhc-vendor-suboption-00.txt because the existing DHCPv6 Vendor
options can be used by the Relay Agent just fine. Again, it is WHERE these
options appear that indicates whether they are for the client (in the client
part of the message) or the relay, in the Relay-Forw or Relay-Reply part of
the message. That work is already done and is not needed!!

- Bernie

-----Original Message-----
From: Wing Cheong Lau [mailto:lau@qualcomm.com] 
Sent: Monday, October 25, 2004 12:50 PM
To: dhcwg@ietf.org
Cc: Ted.Lemon@nominum.com; rdroms@cisco.com; volz@cisco.com
Subject: Fwd: RE: [dhcwg] New draft on DHCPv6 Relay Information Option and
RADIUS Attributes sub-option

Dear Bernie and Ted,

Thanks for your suggestion. My response is interlaced below.

X-BrightmailFiltered: true
From: "Bernie Volz" <volz@cisco.com>
To: "'Wing Cheong Lau'" <lau@qualcomm.com>, <dhcwg@ietf.org>
Cc: "'Ralph Droms'" <rdroms@cisco.com>
Subject: RE: [dhcwg] New draft on DHCPv6 Relay Information Option and RADIUS
Attributes sub-option
Date: Fri, 22 Oct 2004 17:41:17 -0400
Organization: Cisco
X-Mailer: Microsoft Outlook, Build 10.0.6626
Importance: Normal
X-PMX-Version: 4.6.0.99824, Antispam-Core: 4.6.0.97340, Antispam-Data:
2004.10.22.1
X-OriginalArrivalTime: 22 Oct 2004 21:42:20.0795 (UTC)
FILETIME=[025954B0:01C4B880]

A very basic question ... why have a Relay Agent Information option and have
sub-options inside this? And, especially 8-bit suboptions.

It seems to me that with the larger 16-bit DHCPv6 option space, we'd just
define an option to carry the "Radius Attributes" instead of placing this
under a general "Relay Agent" option. The context of the message is pretty
clear -- if it is from the relay, it can only be in a Relay-Forw (and
Relay-Reply) message option area.

There's also another advantage to this. Take the DHCPv4 subnet-selection
option - there are two forms of this, one in for the relay agent and another
for the client. This won't be necessary in DHCPv6 as the LOCATION of the
message (either in the Client's Solicit, Request, ... or in a
Relay-Forw/Relay-Reply) tells us who added it and which we'd prefer to use.
This means only ONE option number is needed.

So, I'd much rather see this specify a base option,
OPTION-RADIUS-ATTRIBUTES, and have this contain the Radius attributes in the
standard Radius encoding. So, it is just 16-bit option code, 16-bit length,
followed by the radius encoding (as 8-bit suboptions).

You can then specify that OPTION-RADIUS-ATTRIBUTES can only appear in the
Relay-Forw (and Relay-Reply) message and MUST NOT appear in client messages
themselves.

- Bernie

The original intent of specifying the sub-option was to try to simplify the
definition of other sub-options in the future, 
e.g. the DHCPv6 version of draft-ietf-dhc-vendor-suboption-00.txt and to
save a byte of so if the relay needs to include multiple sub-options at the
same time.  
On 2nd thought, this "uncommon case"  may not worth the complexity. 

Your suggestion looks cleaner. 
Also, the use of 16-bit instead of 8-bit length field can  eliminate the
255-byte length limit of RADIUS attributes to be included.
I will make it a single-level option according your suggestion in the next
revision.

About the naming of the option, how about
"OPTION-RELAYED-RADIUS-ATTRIBUTES"
instead of "OPTION-RADIUS-ATTRIBUTES" ? This  is to highlight the nature of
this option,
i.e.  being inserted by the Relay Agent.

Thanks again.

Regards,

Wing

-----Original Message-----

From: dhcwg-bounces@ietf.org [mailto:dhcwg-bounces@ietf.org] On Behalf Of
Wing Cheong Lau

Sent: Friday, October 22, 2004 12:45 PM

To: dhcwg@ietf.org

Subject: [dhcwg] New draft on DHCPv6 Relay Information Option and RADIUS
Attributes sub-option

Dear all,

We have submitted a new draft on DHCPv6 Relay Information Option and RADIUS
Attributes sub-option last week. It's already available from the ietf site 

http://www.ietf.org/internet-drafts/draft-droms-dhc-v6-relayopt-00.txt

but I have not seen the official announcement so far.

The draft basically carries over similar capabilities, namely, 

RFC 3046 and draft-ietf-dhc-agentopt-radius-08.txt

from DHCPv4 to DHCPv6, with initial use cases targetting for 

the 3GPP2 environment.

Comments are welcome.

Regards,

Wing

Abstract 

   This document introduces the capabilities of the DHCPv4 Relay Agent 

   Information Option in RFC 3046 and the corresponding RADIUS-

   Attributes Sub-option to DHCPv6. In particular, the document 

   describes a new DHCPv6 option called the Relay Agent Information 

   option which extends the set of DHCPv6 options as defined in RFC 3315 

   and 3376. Following its DHCPv4 counterpart as defined in RFC 3046, 

   the new option is inserted by the DHCPv6 relay agent when forwarding 

   client-originated DHCPv6 packets to a DHCPv6 server. Servers 

   recognizing the Relay Agent Information option may use the 

   information to implement IP address or other parameter assignment 

   policies.  The DHCP Server echoes the option back verbatim to the 

   relay agent in server-to-client replies, and the relay agent strips 

   the option before forwarding the reply to the client. The Relay Agent 

   Information option is organized as a single DHCPv6 option that 

   contains one or more "sub-options" that convey information known by 

   the relay agent.  A RADIUS Attributes Sub-option, following its 

   DHCPv4 counterpart, is also defined.  

>Ted:
>
>Not sure what you're referring to.
>
>A relay would normally take a client message and place it into a Relay-Forw
>with a Relay Message option containing the client's message. It can also
add
>any other options that it wants (and are allowed). In the base spec, this
>might include an Interface-ID option, for example. And, why I suggested was
>that this new I-D just define a "Radius Attribute" option which the Relay
>Agent would add into the Relay-Forw message.

>- Bernie

> -----Original Message-----
> From: Ted Lemon [mailto:mellon@nominum.com] 
> Sent: Friday, October 22, 2004 8:18 PM
> To: Bernie Volz
> Cc: dhcwg@ietf.org; 'Ralph Droms'; 'Wing Cheong Lau'
> Subject: Re: [dhcwg] New draft on DHCPv6 Relay Information 
> Option and RADIUS Attributes sub-option
> 
> 
> Er, furthermore there's already a relay encapsulation method 
> defined in 
> the base DHCPv6 specification.
> 