> In principal, agreed - however, a standardized way of driving that 
> (e.g. via NETCONF) would be quite useful (and we are not there yet),
>  but not in scope here.

Agreed...

> The goal of I2RS should be to support an external interface to the 
> RIB process of a routing element.  If a system designer wants to use
>  I2RS internally within a single routing element instance, that would
>  be valid, but is not the target of the protocol.  My feeling is that
>  most routing system developers/platforms already have that link 
> sorted out :)

I think this is an important/interesting point... Thanks for bringing it
up.

> I do believe that it would be useful for I2RS to get the full 
> candidate database from the various protocols (maybe as they feed in
>  from the routing processes), before best path winnowing is done.  I
>  assume that is in scope here, but is not completely clear.

Yes, I think it is.

_______________________________________________
i2rs mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/i2rs

Reply via email to