[CCAMP] my review of draft-izh-ccamp-flexe-fwk

[Daniele Ceccarelli <[e-mail]>]

Authors,

 

I reviewed draft-izh-ccamp-flexe-fwk. Please find below my comments.

 

*	Figure 1: I don't understand the figure. So far the main topic is
the fact that MAC rate and PHY rate do not necessarily map 1:1 but now I
don't understand what this reference model means. Can you add some text?
*	Section 4: you jump directly to the requirement without describing
what the framework is. I would expect a section before section 4 that
describes what you want to achieve, and then requirements. The direction
we're having lately from the IESG is that requirements not necessarily need
to be published. You can take them out in a separate document that we can
adopt but then not publish or keep in a section of this document and the we
put them in appendix, as you prefer.
*	- You start talking directly about the requirement of supporting the
creation of a flexE group and the only description of a flexE groups is 3
lines in the terminology. The section described above is really needed.
*	Section 5. As above. This needs to be introduced somehow. Moreover I
agree with the creation of the FlexE group, but does GMPLS configure
clients?
*	"Set up of a higher layer LSP that requires to be run over a FlexE
infrastructure."this is something that already exists. Whether an LSP is
created over a FlexE, an optical trail or a classical Eth link is
independent on the underlay technology.
*	"5.1.  Types of LSPs in a FlexE capable network. " isnt' the FlexE
the underlay with respect to the LSP? You might talk about procedures for
setting up LSPs in a FlexE network, not types of LSP.
*	Moreover if FlexE groups and clients are already preconfigured we're
in a classical non FlexE scenario, right?
*	Instead of starting straight with a multilayer control plane I would
talk about just the control plane of FlexE and the eventually talk about all
the possible solutions for the combination of control planes (FlexE and
LSP), confederated, multilayer etc.
*	Probably section 6 is a good candidate for the intro section missing
before section 4 (the one I was describing above)
*	6.2.1 I would spend more time describing the architecture rather
than jumping directly to signaling and routing. Describe the data plane and
what you want to achieve, before jumping into what needs to be advertised
and what needs to be signaled.
*	Section 7. What is a flexE LSP? I expect to read this in the new
section (architecture/framework?)

 

BR

Daniele