[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Failure definition w.r.t. SONET/SDH and Protection/Restoration




Hi Adrian,
                  I see your points (that doesn't mean I agree ;-)).

Given that seems that I'm the only one to ask such kind of statement I think the discussion could stop here.

Regards and thanks

Diego

Please respond to "Adrian Farrel" <adrian@olddog.co.uk>

To:        "Diego Caviglia" <Diego.Caviglia@marconi.com>
cc:        <ccamp@ops.ietf.org>, <Sasha@AXERRA.com>

Subject:        Re: Failure definition w.r.t. SONET/SDH and Protection/Restoration



Hi Diego,
 
>> - The GMPLS control plane does not  necessarily want to define data
>>   plane failures. Rather, it  wants to signal (report) errors and allow
>>   the control plane  state (and data plane state) to remain in place.
> [dc] Agreed and that  was my initial point.  Which kind of data plane (e.g.
> SDG)  defects have to advertised to control plane to trigger restoration?   
> Given that control plane is concerned in  restore the data plane it should
> also define what is a failure.  It is a  fiber break or we have also to
> consider other defects? Of course we  have.  What happens if one
> restoration TNE consider as failure  only Loss Of Light (LOS) while the
> other starts restoration even if there  is a e.g. Loss Of Pointer?  Is
> that acceptable?
 
Let's look at two things.
 
First, the trigger for protection switchover. If  you look at section 6.2 of draft-ietf-ccamp-gmpls-recovery-e2e-signaling-02.txt  (currently in working group last call, so please comment now!) you will  see...
 
   To co-ordinate the switchover  between end-points, an end-to-end
   switchover request is needed  since a failure affecting one the LSPs
   results in both  end-points switching to the other LSP (resulting in
   receiving  traffic from the other LSP) in their respective
   directions.  This is done using the Notify message with a new Error
   Code  indicating "Working LSP Failure (Switchover Request)".

This clearly indicates that the objective is not  to report a failure, but to request a switchover.
 
Which takes us to your second point: which data  plane errors should be reported to the control plane to trigger switchover  signaling, and would it matter if different transport nodes applied different  policies?
 
My contention is that this is entirely an  implementation issue. I think it would be a wise implementation that included  some degree of configurability to this function, but that is not in CCAMP's  scope. If two "TNEs" have different trigger thresholds I do not see that any  harm will be done, it might even be desirable.
 
>> - Decisions that trigger control plane  protection actions are
>>   fundamentally implementation  choices.
> [dc] Here we are.  This is the point.  IMHO CCAMP  should provide some
> guidance on this.
>
>>   Maybe  we should make a statement that this type
>>   of thing SHOULD be  configurable so that there is some sort of
>>   consistency within  a network.
> [dc] Something like "..control plane restoration mechanism  SHOULD be
> triggered as the SNC/I defined in ITU-T G.783 (section  12.1.1.1).
 
But why?
 
It is not enough to know that this is a difficult  question for an implementaiton to answer.
 
Why is it not enough to say that the Data Plane  should trigger control plane recovery action when the data plane deems that  there is a fault from which it determines that recovery is required and from  which it is unable or unwilling to recover on its own?
 
If you can show that there is a need to make a more definitive statement  then we can consider it.

Cheers,
Adrian