| Thread Links | Date Links | ||||
|---|---|---|---|---|---|
| Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
Glen, Given the added complexity of such wavelength-specific switching as well as the utter lack of actual EPON deployments with protection in production, I just have to ask – why are we trying to do more than the minimum to have the same features in Nx25G-EPON as we did 1G/10G-EPON systems? Is there any documented requirement for path selective failover? The one challenge that comes to mind, immediately, is the different delay on different wavelengths traversing different trees, as an example. Even with a common trunk, we can tolerate some skew but what if the construction team messes the design and you have suddenly 10km of difference between one trunk and another? With a simple on/off protection, we have everything we need already – whenever at least one channel detects LoS, we switch to the alternative path, the way we do in 1G/10G-EPON specs. All we need is to add a statement about just that, specifying that any switching operation happens across all wavelength channels present in the given transceiver. With that said, as far as 1) is concerned, I would treat them exactly the same, i.e., trigger the switchover. As far as 2) is concerned, I will again ask for the driver / requirement for this mechanism. Is there anybody deploying anything like that? I cannot find any references, outside of a few research papers focusing on “what could be if … “. I do not think we need to support every single protection scheme under the sun, especially when there is a very limited market driver for them to begin with. I am struggling to remember whether there was a specific agenda for the meeting tomorrow or whether it was just a free for all discussion on protection and power saving mechanisms. As far as power protection is concerned, I have been swamped with production issues and additionally I am sick this week so my ability to focus on power saving I was planning to work this week will be limited Regards Marek From: Glen Kramer <glen.kramer@xxxxxxxxxxxx> Hi Marek, Regarding your last bullet, the ONU receiver specifications in 802.3ca do include the signal detection threshold values, just like the OLT receiver does. See tables 141-21 and 141-22 The main reason why the optical protection was listed as the big ticket item is that we now have multiple channels. We need to make decisions on the following questions:
Trunk protection, ONU detected LoS on one of two channels Trunk protection, OLT detected LoS on one of two channels Tree protection, ONU detected LoS on one of two channels Tree protection, OLT detected LoS on one of two channels Note that 802.3ca CCP includes a message designed to inform the OLT of a channel failure at the ONU
Is this in the tomorrow’s consensus call agenda or the next week? Glen From: Marek Hajduczenia <mxhajduczenia@xxxxxxxxx> Good morning, 1904.4 TF members, Per discussion on the call last night, I started looking at subclause 9.3 protection to see what needs to change to accommodate Nx25G-EPON protection. I made a few changes tracked in the attached document, mostly to fix any issues identified in the document (editorial) as well as align with the IEEE Std 802.3ca text as published. Below is the outline of my major changes / observations Most (if not all) Clause 9 cross references are broken. It seems that the 2016 edition of 1904.1 was broken during the publication process, where cross-references cannot be automatically updated any more. This also resulted in one of tree protection schemes becoming unassociated with any profile in 1904.1. In 1904.4, I suggest that section to be altogether deleted, since it was originally associated with Profile B in 1904.1. I updated MPCP primitives to match definitions provided in 802.3ca – I believe this will make the text cleaner. There is no Signal Detect Threshold defined for ONU receivers in 802.3ca – I assume we will just drop the Optical LoS tracking requirement for the ONU side, since we do not have now anything to point back to in the physical layer spec. Please let me know if there is anything unclear. As it is, I believe subclause 9.3 could be mostly reused with minimum development effort, but requiring the multi-channel Nx25G-EPON system to behave as a single wavelength system. Regards Marek To unsubscribe from the STDS-1904-4-TF list, click the following link: https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-1904-4-TF&A=1 To unsubscribe from the STDS-1904-4-TF list, click the following link: https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-1904-4-TF&A=1 |