Re: [dhcwg] I-D Action: draft-ietf-dhc-dhcpv6-client-link-layer-addr-opt-00.txt

[Simon Hobson <[e-mail]>]

perl-list wrote:

>>Okay, consider prefix-delegation, for example. ...

>What you say in theory is interesting and could pose a problem.  In 
>practice, however, it seems that prefix delegation and WAN 
>configuration happen simultaneously in the same exchange within the 
>same packet.  Also, as this draft is intended to 
>help correlation between DHCPv4 and DHCPv6 using the (hardware / 
>link layer / MAC - whatever we are calling it these days) address, 
>your argument about downstream interfaces that would not be known in 
>DHCPv4 space is irrelevant.

I agree, and what's more I can't really see how it's an upstream 
server's job to configure downstream interfaces on a router. In the 
general case where PD is likely, it's going to be an ISP delivering a 
prefix to a customer's router - what the customer wants to do with 
that isn't really anything to do with the ISP.
All the ISP needs to do is a) configure the customer's WAN, and b) 
provide them with a PD.

While I could see some situations where going further might be 
useful, I could only see it working in a managed environment (eg ISP 
supplied routers, I'm thinking something similar to ISP supplied 
cable modems). I'd see that sort of situation managed by 'private' 
extensions/options which could be agreed between the client and 
server - possibly as an industry segment standard vendor option space.


I'm inclined to agree with Ted - keep it simple so it can actually get passed !

-- 
Simon Hobson

Visit http://www.magpiesnestpublishing.co.uk/ for books by acclaimed
author Gladys Hobson. Novels - poetry - short stories - ideal as
Christmas stocking fillers. Some available as e-books.