Hi, Julien and Young, 

I fully agree on that we should try our best on reusing the existing 
object/TLVs. FYI, when we are working on GMPLS extensions from fixed-grid 
(WSON) to flexi-grid, we have some TLVs in parallel (switching types) for 
fixed/flexi-grid. 

I assume this is exactly what we are doing in this draft, the split occurs when 
it is difficult to use existing TLV to represent the new features. In some 
cases, the meaning of flags/TLVs need to be changed to support different 
scenarios, so the similarity on object does not necessarily mean the equivalent 
on functionality. 

My 2 cents, 
Haomian

-----邮件原件-----
发件人: Pce [mailto:[email protected]] 代表 Leeyoung
发送时间: 2017年7月20日 16:41
收件人: Julien Meuric; [email protected]; [email protected]
主题: Re: [Pce] My Comment on draft-lee-pce-flexible-grid

Hi Julien,

I agree. A New flag on the object header (WA Object, I assume that is what you 
are pointing to) where we have a flag to indicate if this is fixed (WSON) or 
flexi-grid is reasonable instead of creating a new object for a new Spectrum 
Assignment Object.

TLVs require a bit different encoding due to the nature of additional 
parameters for flexi-grid. So strict re-use of WSON TLV may not be sufficient 
for some cases in flexi-grid. 

Best regards,
Young

-----Original Message-----
From: Julien Meuric [mailto:[email protected]] 
Sent: Thursday, July 20, 2017 10:28 AM
To: [email protected]; [email protected]
Subject: My Comment on draft-lee-pce-flexible-grid

Hi,

The discussion during the meeting suggests that I need to clarify my comment 
about draft-lee-pce-flexible-grid.

This I-D is very similar to draft-ietf-pce-wson-rwa-ext, which addresses the 
exact same problem over a slightly different WDM label space (running a 
side-by-side diff between them appears to be very practical).
This new I-D requests the creation of one object and 3 TLVs, which are 
identical to the ones created in the WSON.
As a result, I believe the latter should be reused as a starting popint.
Covering the flexi-grid case may just need to allocate new flag in the object 
header to identify the WDM type we're dealing with, and document the 
flexi-grid-specific assumptions (if any).

Thanks,

Julien

_______________________________________________
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