> It took me a while to figure out that different types of information are 
> being manipulated at each arrow. The diagram would be easier to understand if 
> each arrow was labeled with the type of object which is being inserted or 
> retrieved at each interface:
> 
> A: Local routing process state
> B: Routing process to RIB process export policy rules
> C: Routes in the RIB
> D: RIB process import policy rules
> E: RIB process to FIB best route selection rules

Yes --thanks... I think it is probably important to make certain we not
only define the type of operation at each point, but also the data being
injected/received, at least broadly... The point in building the data
model would then be to define these data transfers more specifically,
and then the protocol work to figure out how to carry the data
specifically called for in an intelligible way, I think...

But I'm thinking way ahead of the ball at this point, and just throwing
out some possible lines of approach. Nothing definite here!

:-)

Russ

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

Reply via email to