[rbridge] Configuration of VLANs vs selfconfiguration

[Guillermo Ibáñez]

Hi Vishwas,
         Yes, regarding the trunking ports, and no, as far as I know, 
regarding the terminal ports, that require specific configuration.  
(Unless the NIC at the hosts are GVRP-aware and preconfigured with the 
VLAN they belong to, and VLAN belonging is distributed in the network 
via GVRP). But this translates the configuration work to the each host., 
IMHO.
In the case of per-port VLANs,  somehow it must be stated which port of 
each bridge belongs to which VLAN.  Actually, this seems to be somewhat 
contradictory with real selfconfiguration (i.e. we split the physical 
network manually in different virtual networks, overriding 
selfconfiguration).
 In the case of per-MAC (dynamic) VLAN belonging, this is something to 
be configured somewhere  at central level (in  the vlan policy server)  
and the result (VLAN belonging), loaded into the rbridge.
      So perhaps we should forget about zero configuration, as long as 
VLANs  shall  still be specified and configured in the campus networks 
to provide separated broadcast domains. Or, alternatively, specify "zero 
IP configuration" in the design objectives.
Regards
GI

Vishwas Manral wrote:

>Hi Guillermo,
>
>Doesn't using GVRP help minimize configuration exchanges in the trunking ports case at least currently? Can't we use that in the Rbridge case too?
>
>Thanks,
>Vishwas
>-----Original Message-----
>From: rbridge-bounces at postel.org [mailto:rbridge-bounces at postel.org] On Behalf Of Guillermo Ibáñez
>Sent: Thursday, September 22, 2005 2:30 PM
>To: Developing a hybrid router/bridge.
>Subject: [rbridge] Configuration of VLANs vs selfconfiguration
>
>
>
>G. Ibáñez wrote:
>
>I have a doubt since some time on VLAN configuration. One of the main 
>aspects driving the routing bridges proposal is to be configuration 
>free, like the transparent bridges, opposite to using  routers where IP 
>addresses must be  asigned, administered and may change when the host 
>moves in the network. However, VLAN configuration of bridges is an 
>important aspect and effort, that prevents selfconfiguration of rbridges 
>to be achieved. In the Rbridges draft (see extract below), this problem 
>is skipped assuming the Rbridges are already configured with the 
>information on which VLANs should be supported on which port.
>Although the problem of VLAN configuration is complex to handle, by 
>ignoring it we risk not to fulfill  one of the main requirements of 
>Rbridges: zero configuration.
>Regards
>Guillermo
>
>Extracted from draft:
>
>"Support of VLANs does traditionally require configuration 
>   of the bridges (or in this case RBridges) to know which links belong 
>   to which VLANs. In theory some other mechanism might allow an RBridge 
>   to know which VLANs should be supported on which port. The RBridge 
>   design does not care how RBridges discover which VLANs are supported 
>   by each of their ports, but for simplicity we assume here that 
>   RBridges (like bridges) are configured with this information." 
>
>
>_______________________________________________
>rbridge mailing list
>rbridge at postel.org
>http://www.postel.org/mailman/listinfo/rbridge
>
>  
>