Lizhong,
 
> [Lizhong] I mean the LAG between two end-points, 
> not MC-LAG. E.g, one LAG interface with physical link 
> A & B between two end-points, it is possible that the forward 
> path will go through A, and backward path will go through B, 
> then the physical path is not symmetric. If I understand correctly, 
> you just assume the transport delay on link A & B are the 

On an ordinary lag, unless it's a MC-LAG this can be assumed to be more or less 
the same.

> same, but this is not always true, especially for the composite 
> link defined in RTGWG (draft-ietf-rtgwg-cl-requirement). 

I quickly glanced through this draft and these can imo be easily avoided by 
defining a new link type - Composite link, in the link type sub-TLV of Link TLV 
in OSPF [RFC3630]. Such links MUST be avoided for setting up the PTP LSPs. 
Similarly the extended IS reachability TLV could be extended for IS-IS to avoid 
composite links.

One could also use coloring to avoid these links, but this would entail manual 
configuration which may not be desirable.

Cheers, Manav
_______________________________________________
TICTOC mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/tictoc

Reply via email to