RE: [RPRWG] RPR and WDM topology discovery

["Vish Nandlall" <vnandlal@xxxxxxxxxxxxxxxxxx>]

Title: RPR and WDM topology discovery

Paul,

 

I had originally asked the question thinking there was context in terms of previous discussion. Your clarification reflects the intent of my question. OADMs typically use a supervisory channel which is extracted at each NE for maintenance (instead of an O-E at each wavelength for the G.709 digital wrapper or Sonet OH). I'm unsure what decision has been made regarding the degree to which the 802.17 MAC is agnostic to the underlying layers, but some thought should be put into OA&M coherency.

Vish Nandlall

Nortel Networks

ESN 393-1350
External (613) 763-1350
vnandlal@xxxxxxxxxxxxxxxxxx

-----Original Message-----
From: Paul Obeda [mailto:pobeda@xxxxxxxxxxxxxxxxxx]
Sent: Thursday, August 09, 2001 10:40 AM
To: stds-802-17
Cc: bjlee; Nandlall, Vish [CAR:KCQ0:EXCH]
Subject: FW: [RPRWG] RPR and WDM topology discovery

Greetings,

 

I'm going to try to reset the context for the question here; if I get that right (or close), then we can retry the underlying question.

 

As I read it, the question addresses the relationship between the 802.17 MAC and a particular underlying PHY.  In this particular case, I believe that the hypothetical stack is RPR MAC (native) over WDM physical layer, with limited OAM capabilities (particuarly for fault sectionalization) in that WDM system.

 

To try to set the context further, I'd like to consider (just for a moment) a few alternative hypothetical PHYs.

 

Let's start with a RPR which happens to use SONET payload formating, but doesn't make use of any of the SONET overheads.  This particular ring doesn't interact with any other SONET gear, and the entire infrastructure it uses is dedicated to that RPR. 

 

Next, let's consider an RPR which exists within a SONET ring, as a closed cascade of bidirectional SONET Paths.  The RPR functionality is (probably?) the same, but the underlying PHY provides a different set of OAM capabilities to the network.

 

We can even consider having multiple independent RPRs on that same SONET gear, if we'd like.

 

Then we can consider multiple SONET subnetworks operating independently over a WDM infrastructure, without any integration between the SONET and WDM layers.  These SONET subnetworks could contain one or more RPRs each (as above), each of which is oblivious to the fact that the WDM infrastructure underneath them is helping them share fiber.

 

If we'd like, we can also consider other layers (such as the Optical Transport Network being pursued by ITU-T SG 15) as either present or absent from the stack.

 

Really, any of those (or other) physical layer technologies may be present or absent underneath the RPR.

 

 

If this is the correct context for the question, then I believe the question becomes one of asking what the division-of-responsibilities is / should be between the 802.17 MAC and any underlying PHY, and futher, what (if any) signaling mechanisms must exist between the MAC and the PHY.

 

This issue probably deserves 802.17 consideration as part of the OAM&P work.  Any comment? 

  

Regards,

 

Paul Obeda

 

-----Original Message-----
From: owner-stds-802-17@xxxxxxxxxxxxxxxxxx [mailto:owner-stds-802-17@xxxxxxxxxxxxxxxxxx]On Behalf Of Vish Nandlall
Sent: Tuesday, August 07, 2001 4:38 PM
To: 'B J Lee'; stds-802-17
Subject: RE: [RPRWG] RPR and WDM topology discovery

see below

 

Vish Nandlall

Nortel Networks

ESN 393-1350
External (613) 763-1350
vnandlal@xxxxxxxxxxxxxxxxxx

-----Original Message-----
From: B J Lee [mailto:bjlee@xxxxxxxxxxxxxxxxxx]
Sent: Tuesday, August 07, 2001 3:57 PM
To: Nandlall, Vish [CAR:KCQ0:EXCH]; stds-802-17
Subject: RE: [RPRWG] RPR and WDM topology discovery

I do not think we identifed such a requirement so far, if I understand your question correctly.

Do you mean that a RPR node supports multiple rings on different wavelengths including

bypasses, acting as an interconnection attachment?
[Nandlall, Vish [CAR:SB06:EXCH]] 

 

I didn't necessarily mean multiple rings supported on one RPR node (although it could be), I was thinking multiple RPR rings running on the same physical plant. One RPR node would terminate and source a wavelength dedicated to its ring, and optically bypass all other wavelengths. In this scenario, it would be difficult to sectionalize a fault if the nearest physical adjacent node was on a different wavelength.

 

 

It would be appreciated if you could clarify more on  

 "will RPR only connect as a client to an existing DWDM network and rely on whatever

  existing transport mechanism is in place?"
[Nandlall, Vish [CAR:SB06:EXCH]] 

 I was referring to a configuration whereby the RPR node subtended an OADM box to access a wavelength in a DWDM ring (which in turn would provide connectivity to other RPR nodes hanging off upstream/downstream OADMs on the same wavelength). Multiple RPR rings could then be realized using the same physical ring structure. This would obviously require the development of a new client interface on the OADM to interface the RPR.

 

 

 

BJ 

-----Original Message-----
From: owner-stds-802-17@xxxxxxxxxxxxxxxxxx [mailto:owner-stds-802-17@xxxxxxxxxxxxxxxxxx]On Behalf Of Vish Nandlall
Sent: Tuesday, August 07, 2001 1:06 PM
To: 'stds-802-17@xxxxxxxx'
Subject: [RPRWG] RPR and WDM topology discovery

In the context of an DWDM RPR implementing optical bypass, would there be a requirement for a dedicated supervisory channel for topology discovery (i.e. if both neighbors of an RPR node add/drop a different wavelength then there is no in-band mechanism for detecting physical adjacencies) - or will RPR only connect as a client to an existing DWDM network and rely on whatever existing transport mechanism is in place ?

Regards,

Vish Nandlall

Nortel Networks

ESN 393-1350
External (613) 763-1350
vnandlal@xxxxxxxxxxxxxxxxxx