[RPRWG] RE: Questions on IPS
Barnali,
I'm cc.ing the 802.17 mailing list on the response, since
I think someone there will have the ability to correct me
if I say something that's wrong.
-Anoop
> -----Original Message-----
> From: Barnali [mailto:barnali.jesudas@xxxxxxxxx]
> Sent: Thursday, February 07, 2002 2:51 AM
> To: anoop@xxxxxxxxxxxxxx
> Subject: Questions on IPS
>
> I have been going thru' darwin draft. I have few questions on
> IPS wrapping is selected. would be thankful for your response.
The term IPS is not use anywhere in the standard. But I'll
answer the questions based on my understanding of the
Darwin proposal.
> Q1) Do the LongPath messages that are
> sourced from a wrapped node continue until the failure
> is cleared? OR Do they stop after the protection is
> completed ( both nodes wrapped ) ?
As far as I can tell, the message is sent only once. The
protection messages are needed only for steering (when wrapping
only the local nodes need to detect the failure). Also,
per the draft, eventually, the topology discovery algorithm
will find out about the failure and all nodes will stop
forwarding traffic that would need to use the failed span.
> Q2) Consider a ring set up as shown in the following figure:
>
> ---------------------------------------
> | ----------------------------------- |
> | | | |
> | v | v
> ----- -----
> | A | | B |
> | | | |
> ----- -----
> ^ | ^ |
> o | | i | |
> u | | n | |
> t | | n | |
> e | | e | |
> r | | r | |
> | v | v
> ----- -----
> | D | | C |
> | | | |
> ----- -----
> | | | |
> | | | |
> | ----------------------------------- |
> ---------------------------------------
> Case 1:
> 1. There is a SF on node A between A and B
> 2. After sometime FS is applied on node D between C and D.
>
> Will the ring be broken in 2 parts, one having A (SF) and D(FS), and
> the other with B (SF) and C (FS)?
That appears to be the case. So you end up with 2 islands.
>
> Case 2:
> 1. There is SF on node A between A and B.
> 2. Next FS is applied on A between A and D.
> Will node A get isolated because of wrapped on both side?
Seems so.
>
> Case 3:
> 1. There is FS on node A between A and B.
> 2. Next SF is detected on A between A and D. ( Order of condition
> has changed from Case 2).
> Will node A get isolated because of wrapped on both side?
Again, seems like that's what'll happen.