Re: Re: [dhcwg] WG Last Call for DHCP draft

On Mon, 20 Dec 2004, Bill Strahm wrote:

> Vivek,
> 
> One question is hard, one is easy...
> 
> On Wed, 2004-12-15 at 20:55 -0700, Ted Lemon wrote:
> > This draft uses a SHOULD to specify what data should be in the client 
> > identifier. This is in conflict with another draft that's in the 
> > process of working its way through to last call - 
> > draft-ietf-dhc-3315id-for-v4-??.txt. This is a bad thing - someone 
> > reading this spec will wonder which of the two specifications 
> > supersedes the other. I don't see any reason to specify the format of 
> > the client identifier in this draft - instead, refer the reader to 
> > draft-ietf-dhc-3315id-for-v4-??.txt and to RFC2132 for advice on how to 
> > format the client-identifier option.
> > 
> This is the hard question... What should we do about this... I think we
> should take this  request - and say our client ID SHOULD follow RFC2132
> (if I understand the problem correctly)... Is there any non-capped
> advise that we can give ?
> 
> Can you propose a change for this Vivek ?

I have not had a chance to go over the draft Ted pointed to -- it appears to
be version -03.txt. 

However, the DHCP over InfiniBand draft does not define the client identifier 
option format itself but only describes the choices for client-identifier 
field itself. These choices are InfiniBand specific and are not likely to
be covered by a generic draft.

The text can be further clarified that the "SHOULD" only applies to the 
selection of one of the formats as described in RFC2132 to avoid any
possible confusion.  

The content of the client-identifier under Type 0 as described in the DHCP 
over InfiniBand draft is conformant to RFC2132. So is the case with the use 
of Type 32. 


> 
> > Also, you have a reference to RFC1542. This is superseded by RFC2132, 
> > so you should refer to RFC2132 instead - it is a huge mistake for 
> > someone to read RFC1542 as advice on how to implement a DHCP client, 
> > and so this draft shouldn't imply that the reader ought to look at 
> > RFC1542 - he or she oughtn't.
> > 
> OK - this is the easy one.  Can you change the reference to RFC2132.

I believe both references exist in the current draft. I'll drop the 1542 
reference.

> 
> Any objections to this for closing this issue ?
> 
> Bill
> > Ralph, what's going on with draft-ietf-dhc-3315id-for-v4-??.txt? This 
> > is getting a little bit embarrassing... :'}
> > 
> I think this is DHC work - nothing for us to see here - IPoIB is moving
> on
> 
> Bill
> 
> 
> _______________________________________________
> IPoverIB mailing list
> [email protected]
> https://www1.ietf.org/mailman/listinfo/ipoverib
> 
>