Hi,
   Thank you for the reply. 
IMPO there is a need for such a functionality: if after all the PCEs processing 
work  we have to rely on crankback, then the question is why do we use PCEs at 
the first place. There may also be other use cases that can benefit from 
prereservation, and we should not only consider this from  a "cranckback 
saving" angle.

It seems that there are 2 distinct points to the discussion regarding the draft:
1- Consensus that there is a problem if resources are not pre-reserved. 
2- There are various ways to solve this problem, each with pros and cons.

I think point #1 is clear enough to anyone who faced or studied the problem. 
And if I recall correctly, the comments after Oscar's presentation (including 
those regarding the NMS-PCE relation) were all related to point #2 above. I 
think these comments can all be addressed by the wg. 

Best regards,
Meral




-----Original Message-----
From: Julien Meuric [mailto:[email protected]] 
Sent: April-11-12 10:39 AM
To: Meral Shirazipour
Cc: JP Vasseur; [email protected]
Subject: Re: [Pce] draft-gonzalezdedios-pce-reservation-state-01 next step ?

Hi Meral.

I believe the freshly posted minutes bring an answer to you question.

Beyond the generic scope in which this I-D was presented, I feel the advantages 
and drawbacks of that proposal are off balance. On the one hand, focusing on a 
single PCE solution removes a lot from the PCE architecture (or requires heavy 
complementary mechanisms, i.e. yet another strong shortcoming). On the other 
hand, it aims at "saving a few crankbacks" (quoting Oscar) in the control 
plane, which is likely to end up by saving a negligible time with respect to 
the data plane timescale, especially for WDM where the label resource is 
scarcer (1st and longer use case in the draft).
Finally, knowing that an implementation could make use of the proposed idea 
based on a simple PCE policy, such an unbalanced protocol extension seems more 
than questionable...

Regards,

Julien


Le 29/03/2012 19:55, Meral Shirazipour a écrit :
>
> Hi JP,
>
>   I may have missed part of the discussion today after Oscar 
> presented. What was the next step for draft:
> http://tools.ietf.org/html/draft-gonzalezdedios-pce-reservation-state-01 ?
>
> Thanks,
>
> Meral
>
>
> _______________________________________________
> Pce mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/pce
_______________________________________________
Pce mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/pce

Reply via email to