nanog mailing list archives

Re: Packet over SONET failback


From: Jason Lixfeld <jason () lixfeld ca>
Date: Mon, 14 Feb 2011 07:53:38 -0500


On 2011-02-14, at 6:47 AM, Rob Evans wrote:

PoS failure detection happens in under 50ms, but what about the failback?  Same deal?  I ask because I've got two 
routers connected to opposite ends of a spare PoS link that I've been playing with and I'm noticing that the 
failback on the far side seems to be about 15 seconds (assuming the near side failover was initiated with an 
interface shutdown command and thusly no shut'd to re-enable the link).

I think there are a couple of issues at play here.  First of all,
SONET/SDH restoration happens at layer 1, whereas it looks like you're
waiting for a router to reroute.  Your reroute times will be tied to
recalculation of IGPs.

Secondly, is this with a Cisco?  Try setting "pos ais-shut" on both
sides.  Unless you do that, the router won't generate and AIS, and it
will take the encapsulation timeout (HDLC, PPP) for the interface on
the other side to go down and signal that to the routing protocols on
top.

It's becoming clear from the responses that my initial post was lacking a bit of info.

It's not actually SONET all the way through.  It's GigE from the router to the SONET node, an unprotected OC192 wave to 
another node, out GigE to the far end router.

I didn't mention that because I didn't think it mattered; I assumed the speed in which the GigE on the other side came 
up was protocol agnostic.  I assumed that as soon as the far end of the SONET node saw the near end go back up, it 
would turn the laser back on on the far side bringing the interface back up.  I assumed also that since it took a whole 
15 seconds that it had something to do with the SONET side of things.

I'm not necessarily talking about forwarding of packets here, I'm talking simply about the time it takes the far side 
interface to come back up over a SONET node; layer 1.  Maybe this has nothing at all to do with SONET, I dunno :)   It 
was gleaned that the next step might be to look at the SONET node and see if it's waiting 15 seconds to turn the laser 
back on or something.

Current thread: