FW: Evaluation: draft-ietf-ccamp-gmpls-architecture - Generalized Multi-Protocol Label Switching Architecture to Proposed Standard

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

Dimitri, if for any reason (possibly on based on other comments
from IESG) you are going to do another rev, can you pls check
the below comments and see if some additional editorial changes 
make sense?

By the way, you can see that Harald also noted the "many Work in Progress"
documents that are listed as normative references.

Thanks,
Bert 

-----Original Message-----
From: Harald Tveit Alvestrand [mailto:harald@alvestrand.no]
Sent: zaterdag 26 april 2003 18:03
To: IESG Secretary; Internet Engineering Steering Group
Subject: Re: Evaluation: draft-ietf-ccamp-gmpls-architecture -
Generalized Multi-Protocol Label Switching Architecture to Proposed
Standard

Harald Alvestrand has no Objection.

note: this does not mean that I like it.....

In particular, I don't like the fact that the document can't decide whether 
it's an architecture, an introduction or a specification. It goes into lots 
of details that are ALMOST specified, but then just says that these are 
"finished elsewhere" - sometimes giving, sometimes not giving, the forward 
pointers.

It's just about right (although overly detailed, and lacking some pointers 
to where the "real spec" is) for an introduction.

It also *contains* an architecture description - the idea of a control 
plane that is a fully connected IP network (with, apparently, manual link 
configuration), a set of links that are not necessarily the control links, 
and in many cases not even IP-capable links, and groupings of these that 
can be manipulated and controlled in various ways, and a set of protocols 
and ways to use those protocols that are defined in other documents.

But the lenght and the level of detail means that it's very hard for me to 
be sure I have grasped that architecture correctly and fully.

And - note - I have said nothing about whether or not I *like* that 
architecture.

Note: It will be a while before this is published as an RFC. There are no 
less than thirteen "works in progress" in its normative references.

                     Harald