The MPLS WG Archive

[David Charlap <David.Charlap@marconi.com>]

Dutta, Pranjal wrote:
> Hi,
>              I think most of the vendors implement min 1 seconds as
> hello-intervals as lesser than that might lead to scalability issues (I
> am not sure if that is the reason why vendors implement at least 1 sec).
> Hardware level detection will be possible only when physical layer has
> end-to-end signaling mechanism for failure. In the case where end-to-end
> signaling mechanism is not possible, the PLR will still see that link is
> up and keep forwarding traffic to downstream LSR, but actually the link
> is failed. I am basically looking for some mechanism that can detect
> "Node Failure" as early as possible and PLR will initiate node
> protection accordingly. I think some kind of OAM mechanism might help.

I think you're misunderstanding the purpose of fast reroute.

Fast reroute is designed to kick in only when the immediate downstream 
neighbor has failed.  Detecting link-level failure in hardware is 
perfectly fine for this.

If the failure takes place elsewhere in the network, then either the 
node immediately upstream of the failure must use FRR, or some other 
mechanism must be used to reroute the traffic.

You would never want to trigger an FRR failover in response to an 
end-to-end detectino mechanism.  One big reason is tha any end-to-end 
system will take so long to detect the failure that there would be 
little advantage to FRR vs. simple switching the traffic to a standby 
LSP (1:1 or 1+1 protection).

Second, an end-to-end mechanism may detect a failure anywhere along the 
path.  If the failed node is not your immediate downstream neighbor, 
your bypass/detour LSPs may not even work to route around the failure. 
And if a downstream node also uses FRR, the multiple simultaneous 
attempts at local-repair may well step on each other, making the 
situation worse.

If you want your originating node to detect the failure with an 
end-to-end mechanism and then signal the node upstream of the failure to 
begin FRR, why bother?  The amount of time it would take for this would 
be longer than the time to simply compute a new path and signal a new 
LSP.  (I'm assuming that that node can't detect its downstream 
neighbor's failure directly, because if it could, there would be no need 
for any end-to-end mechanism in the first place.)

-- David

_______________________________________________
mpls mailing list
mpls@lists.ietf.org
https://www1.ietf.org/mailman/listinfo/mpls