FW: draft-ietf-ccamp-gmpls-architecture-05.txt

["Wijnen, Bert (Bert)" <bwijnen@lucent.com>]

While review is taking place in the IESG, we've got 
this comment. Does it warrant a small piece of text 
to be added, or do we believe it has been addressed
properly already?

If text is to be added, then pls send suggested text
in an email with an indication as to where we want to
put it. Do not send a new revision of the ID (yet).

Thanks,
Bert 


At 04:47 AM 4/16/2003 -0700, Randy Bush wrote:
>> one immediate question comes to mind ... they talk of a link
>> management protocl (lmp) (eg in chap 8) it seems to do just
>> about what ppp does. the obvious question is 'why not use
>> ppp?'
>
>ppp control is done in-band. LMP supports the out-of-band mode, which
>is important in optical networks where you can't simultaneously switch
>some things through (user data) and take a look at other (control
>messages). The out-of-band nature places some unique requirements
>(such as 1:n control to data link ratio, link property correlation,
>failure detection) that warrant a separate protocol.

If it isn't already explained in the document, one might want to 
add such an explanation. Documenting _why_ things are done is just
as important as document _what_ is done.