Hi,
The following paragraph is from
draft-ietf-mpls-generalized-signaling-02.txt.
...
Note that since the nodes sending and receiving the new form of label
know what kinds of link they are using, the generalized label does
not contain a type field, instead the nodes are expected to know from
context what type of label to expect.
...
I'm not clear how the label type would be inferred, solely on the
knowledge
of link types as mentioned in the above paragraph, if the link id and
label
type field are not provided alongwith the label when we have the scenario
of two hybrid nodes connected together having the following:
* The nodes have K no. of fibers between them.
* Each fiber carries L no. of wavelengths.
* Some of the wavelengths are SDH/SONET (TDM) switched and
the others are lambda (or wavelength) switched.
* The control channel (for GMPLS control signalling) is OOB/OOF
(Out of Band/Out of fiber)[Referring to LMP].
If the "context" refers to the context based on the identifier of the LSP
(through lightpath/cr-lsp id), then it makes sense.
Besides this, I have one comment. Though the GMPLS Func. Desc. draft
specifies
the format of waveband labels, but the companion documents, like
architecture
document, don't specify in what all scenarios such labels are going to be
used/granted. IMO, it leaves a lot to assume.
Pls. help and correct me if something has been overlooked.
Thanks in advance.
Prem