Sanjay Wadhwa | 30 Jan 07:50 2005

GSMPv3 and L2C

DSL Forum is working on a protocol for interaction between access-nodes(e.g DSLAMs) and BRAS devices
(controllers in gsmp-speak) in an atm or ethernet based access networks.
Avri had posted on the list the DSL forum draft on this work (L2CP - L2 control protocol). The idea is to extend
GSMPv3 to implement requirements of the L2 control protocol. 
In going through the current GSMP RFCs, and the DSL forum document a couple of observations :

1. GSMP RFCs suggests that switch is the passive entity with respect to tcp connectivity. The controller
initiates the TCP connection. In certain cases (e.g in L2C it makes sense to let 
    the controller be passive w.r.t TCP connection, and let the access-node(s) initiate TCP connection). 
Does it make sense for the GSMP spec to not require that controller initiate the tcp connection 
   but keep it be open such that this behaviour can be config driven on the devices involved. The applicability
document can capture specifics of different applications.

2. The adjacency protocol in GSMP has version number negotiation support. Was there any discussion when
GSMP was being designed to take a more general approach to capability negotiation 
    between the swtich and controller (e.g  by including capability TLVs in adjacency protocol messages) ?
This helps extensibility in general.

Will appreciate feedback (including general discussion on extending GSMPv3 to implement L2C as
specified in the DSL forum doc)...