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

On Jan 25, 2005, at 10:50 PM, Vivek Kashyap wrote:
> I've kept the reference to RFC1542 because RFC 2131 itself points to 
> RFC1542
> when discussing the ramifications of using the BROADCAST flag.

I understand now why you want to refer to RFC1542.   I would suggest 
this text instead:

Although [RFC1542] discourages the use of broadcast, in the case of 
IPoIB, broadcast is necessary, since there is no way to respond with a 
unicast.   Readers should be aware that RFC1542 has been superseded by 
RFC2132, and should refer to RFC2132 rather than RFC1542 for guidance.

The point of this change is simply that I don't want a new draft to do 
anything to encourage the reader to think that they need to read 
RFC1542.   RFC1542 contains quite a few errors, which is why it was 
replaced by RFC2132.   We have had problems in the past with people 
reading RFC1541/1542 instead of RFC2131/2132, and the entire reason for 
my objection to your reference to RFC1542 was to avoid encouraging 
this.

Regarding the text about client identifiers, I am extremely 
uncomfortable with this.   Obviously it's not just my decision, but 
right now you haven't given me any reason to think this is a good idea, 
and I've given you some pretty definite reasons why I _don't_ want it.

Pretty much the only way I would be comfortable with any text like this 
is if you first referred the implementor to the 3315id draft and then 
suggested these formats as what to do if for some reason the 
implementor didn't follow the 3315id draft.   Again, my reason for 
saying this is that I don't want it to be the case that someone reading 
this draft would get the impression that it somehow superseded the 
3315id draft.

I would also suggest adding some text saying that these client id 
formats are suggestions, and that DHCP servers MUST NOT consider it an 
error if the client id differs from the suggested format.