| Thread Links | Date Links | ||||
|---|---|---|---|---|---|
| Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
|
Dear all:
A motion was passed at last meeting that results two upstream options for 25G, 1270nm (let’s call it blue 25G) and 1310nm (let’s call it red 25G). It seems a progress on the prolonged
wavelength plan discussion, however, it does NOT solve the problems it claimed. Furthermore, it creates more problems that have much negative impacts on products and networks in the future.
The reason for two 25G upstream options, whether call it brownfield/greenfield or not, to my understanding, rooted with seeking of WDM coexistence with 10G PON (10G EPON, XGSPON) and
GPON. The 1270nm blue 25G supposed to WDM coexist with GPON, and the 1310nm red 25G option WDM coexist with 10G PON.
The immediate drawback, on the surface, is that it creates TWO distinct 25G EPONs with 2 different sets of optics/colors, that divides the optical component market for the 25G EPON.
Other negative impacts are more profound, but I’ll leave them for now and tune to discuss whether it solves the coexist problems it claimed.
Take coexist with GPON for example, GPON bandwidth is already under constrain for offering Gigabit services today, upgrade to 10G PON is imminent or happening, therefore, at the time
for 25G EPON deployment, 10G PON may already widely deployed in field, some coexist with GPON. This gives two coexist scenarios for 25G EPON: a). GPON and 10G PON (XGSPON for example) already coexist on the same ODN; b). GPON and 10G PON are on different
ODNs.
For scenario a, neither blue 25G EPON nor red 25G EPON could coexist with the GPON and XGSPON that already share the ODN.
For scenario b, the blue 25G could coexist with GPON and the red 25G EPON could coexist with 10G PON on separated ONDs. It seems solve the problems at expense of colored 25G, however,
this creates operational problems that exist in the lifespan of the 25G EPON. Considering that if 10G and 25G already deployed, we’ll expect that 2.5G PON will gradually phase out. The net result is that there will be two 25G EPON islands, red and blue, in
operators’ networks, to the end of the 25G EPON lifespan. They are isolated in otherwise physically neighboring networks. It will have much negative impacts on OPEX.
Therefore, the two optioned 25G upstream wavelength plan does not solve the coexist problems it claimed, and on the opposite, it creates outstanding problems on products and network
operations.
There is a simple solution for 25G upstream, only one upstream wavelength is needed to enable coexist with 10G PON and GPON – 1270nm (+/- 10nm). It could WDMA coexist with GPON and TDMA coexist with 10G PONs.
Eugene
From: Glen Kramer <000006d1020766de-dmarc-request@xxxxxxxx>
Sent: Tuesday, August 15, 2017 4:55 PM To: STDS-802-3-NGEPON@xxxxxxxxxxxxxxxxx Subject: Re: [802.3_NGEPON] Call for agenda items -- weekly consensus building for IEEE 802.3ca (100G-EPON) Frank,
The ONUs should neither know or care whether they are deployed in Brownfield or in the Greenfield.
A motion at the last meeting adopted these two plans:
Plan BfX: wavelength plan targeting Brownfield deployments coexisting with 10Gb/s PONs (10G-EPON, XG-PON1, and XGS-PON): 1300-1320 nm for .3ca 10G and 25G upstream (dual-rate) Plan BfG: wavelength plan targeting Brownfield deployments coexisting with GPON: 1260-1280 nm for .3ca 10G and 25G upstream (dual-rate)
All this proposal suggests is that in the Greenfield, the 25G/25G ONUs should use BfG optics and 25G/10G ONUs should use BfX optics.
Also note that PON overhead itself is asymmetric. Upstream has burst-mode overhead, discovery overhead, and a slightly higher FEC overhead due to shortening of last codeword in a burst. So, we do need slightly higher upstream capacity if we want a strictly symmetric goodput. Capacity asymmetry is an operation issue and can be solved by operational means. The OLT does not have to grant each upstream channel to its maximum capacity.
Thank you, -Glen
From: frank effenberger [mailto:frank.effenberger@xxxxxxxxxx]
Use two different downstream wavelengths for the two upstream wavelengths, and you got my vote. I think it is quite clearly indicated in your 25+10 upstream greenfield system. If we don’t use two separate wavelengths, then you have a system with 25G downstream but 35G upstream capacity. And that doesn’t fit the traffic profile anywhere.
Frank E.
From: Harstead, Ed (Nokia - US) [mailto:ed.harstead@xxxxxxxxx]
Marek,
I hope this will clarify:
We passed a motion in last meeting that allows 10G and 25G upstream wavelengths to use either 1270 or 1310 nm (but does not allow the 1270 nm option in the case of 10G EPON ONUs on the same ODN). Therefore we already have two options for .3ac 10G upstream wavelengths, and our new proposal does not add any additional wavelength options. Our new proposal does not add any additional ONU variants either. The only addition is a new OLT variant.
The benefit of the new greenfield-optimized OLT variant is that 10G upstream from 25/10 ONUs and 25G from 25G/25G ONUs do not have to share the same TDM channel. We have certainly heard that operators dislike 10G + 25G sharing.
Hope that helps, Ed
From: Marek Hajduczenia [mailto:mxhajduczenia@xxxxxxxxx]
Ed,
I find discussion on slide 3 confusing, in that 10G US wavelength for 802.3ca devices, per your earlier slides, is around 1310nm, which makes sense, given previous discussions. On slide 3, you seem to imply that we really have two 10G US wavelengths that can be picked and chosen, as needed, depending on whether an operator plans to use 10/10G-EPON at all or not. Seems to me we are adding way too much rope for ourselves down the road, as far as number of different optical modules is concerned
I am OK with your Bfx and Bfg wavelength plans, but I do not see any need for Gf plan and added confusion to the table.
Marek
From: Harstead, Ed (Nokia - US) [mailto:ed.harstead@xxxxxxxxx]
Here is the contribution on 802.3ca 10G US channels in greenfield.
Ed
From: Curtis Knittle [mailto:C.Knittle@xxxxxxxxxxxxx]
Dear Colleagues,
We have at least one contribution for the 100G-EPON (IEEE 802.3ca) consensus building meeting scheduled for this Thursday, August 17, 11:30 am – 1:00 pm MDT.
Please let me know by 5:00 pm MDT Wednesday if you have any other topics for the agenda.
As a reminder from the closing report of our last F2F meeting in Berlin, the following topics were deemed important for contributions:
q Actions: – Par/Objectives/Scope change – Power budget • Close on 25Gbps power budget (Ed/Dekun/Umeda) – MPCP/MPRS • Preliminary proposals for changes to channel bonding scheme (Duane, Glen)(Aug 31) • Discovery process / configuration / ONU capabilities field in RegREQ (Duane/Dan) – Fault Tolerance/Recovery/Loss of Channel – Wavelength plan • 2 versus 1 DS0 (Frank/EdH)(August 17) – SOA preamp/dynamic range • Cost and feasibility still needs to be demonstrated – Line coding (Postponed until after FEC decision) – FEC information • Comparisons of parity matrices and different FEC (Mark/Bill/Yin) – SERDES/CDR • Check margin for 10^-2 input BER (Yin, Ryan) – 50G single wavelength • Technical feasibility of upstream bursts (Critical path item) (Nokia) (likely September) • Wavelength plan • Power budget • Modulation
Thank you,
Curtis
Curtis Knittle VP Wired Technologies – R&D CableLabs desk: +1-303-661-3851 mobile: +1-303-589-6869
Stay up to date with CableLabs: Read the blog and follow us on Twitter
|