The MPLS WG Archive

[Heiles Juergen <Juergen.Heiles@icn.siemens.de>]

Hi,

while reading through draft-ietf-mpls-generalized-signaling-01 "Generalized MPLS - Signaling Functional Description" I noticed that the document is not very general in its description, it is focused on SDH/Sonet and WDM.
Already in the introduction it limits itself to certain multiplex technologies. It subdivides LSRs into
- Packet-Switch Capable PSC (which could be IP, ATM or any other packet/cell based technology)
- Time-Division Multiplex Capable TSC
- Lambda Switch Capable LSC
- and Fiber-Switch Capable FSC.
The later 3 are only specific technologies for circuit switching and should as such not listed as generic sub classes. 
Instead I would subdivide into 
- Packet-Switch Capable PSC
- and Circuit Switch Capable CSC, which could be time-slot switching (TDM with its different flavors like SDH/Sonet and PDH), wavelength/lambda switching (WDM), but also frequency switching (FDM) or code switching (CDM)DM or fiber/wire switching.
The differentiation into LSC ands CSC is important and not if it is TDM, WDM, FDM, CDM, ... This gets important if you go into more details of CSC in the same way as IP or ATM  gets important for LSC.
In draft-bernstein-gmpls-optical-00 "Some Comments on GMPLS and Optical Technologies" some of the specific features of circuit switched networks are listed (e.g. a connection uses a specific amount of link bandwidth, the link has to support the specific format/bandwidth of the connection). They apply to circuit switching in general.
So the first focus of GMPLS should be on the specific features/requirements of PSC and CSC. Additional technology specific features/requirements (e.g. for SDH/Sonet, WDM) should be part of a technology specific definition as it  is not longer general.

Regards

Juergen Heiles