Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: [802.3_NGEPON] 答复: [802.3_NGEPON] Call for agenda items -- weekly consensus building for IEEE 802.3ca (100G-EPON)



All: We have been discussing coexistence somewhat independent of “practical” migration seniors  that results many coexistence  requirements which turns make wavelength allocation too complicated. Complicated  wavelength plans will not only make current project difficulty but also cause operational complications and negative impact future PON standards in the long run.

 

Access network upgrade normally involving CPE or OSP modifications which is very costly. Therefore, when upgrade the access network to a new rate, it is expected to be in operating for many years. The 4X rate increase seems to work well in the past. The 10G to 25G migration seems too small a step. If we only consider “practical” migration paths, the wavelength allocation could be much simpler.

 

Eugene Dai, Ph D

Principle Transport Arcitecture

Cox Communications

Eugene.dai@xxxxxxx

404-269-8014

 

From: Marek Hajduczenia [mailto:mxhajduczenia@xxxxxxxxx]
Sent: Thursday, October 26, 2017 11:18 AM
To: STDS-802-3-NGEPON@xxxxxxxxxxxxxxxxx
Subject: Re: [802.3_NGEPON]
答复: [802.3_NGEPON] Call for agenda items -- weekly consensus building for IEEE 802.3ca (100G-EPON)

 

It is all fine but not all operator will do what you outline. For me, 25G-EPON will be added to existing 10G-EPON deployments where needed to and deployed where there is no PON. I would never rip out 10G-EPON just because there is a newer gen out there. You assumptions work for a specific operator, but not all scenarios …

 

From: Liudekun [mailto:liudekun@xxxxxxxxxx]
Sent: Thursday, October 26, 2017 9:01 AM
To: STDS-802-3-NGEPON@xxxxxxxxxxxxxxxxx
Subject: [802.3_NGEPON]
答复: [802.3_NGEPON] Call for agenda items -- weekly consensus building for IEEE 802.3ca (100G-EPON)

 

All:

I fully agree with frank’s comments. That’s what I plan to answer.

I just want to add a little more.

I strongly supported that :  when we talk about “50G”, it’s not only one system: : it includes both 50G symmetric and  asymmetric.

 

What’s more, based on my observation: asymmetric 50G is not only 50/25, but high likely also includes 50/10G

 

For the residential users , the downstream bandwidth  is much more important than the upstream bandwidth , where for the cost of ONU,  the transmitter is always the major cost (compared with the cost of receiver )

That’s why there is a so big volume on the asymmetric ONUs  (Actually , the asymmetric ONUs are the major volume)

 

For example ,as I know, more than 90% of the 10G EPON ONUs shipments by far are 10G/1G ONUs (the asymmetric 10G ),  similarly ,  I would like to predict 50G/10G will be a quite absorbing system for a lot of operators (due to the 10G transmitter will be really in low cost when 50G is need then, it will be quite suitable as the ONUs’ transmitter )

 

 

A further response to John’s question :

so the NRZ signal may have to be limited to the lower ER that is used for PAM4 (~5dB).”

 

I agree. But this is quite fine. Due to when 25G OLT is upgraded by 25/50G dual rate OLT,  I believe the new OLT transmitter will have a higher launch power than previous 25G OLT (due to the lower sensitivity of 50G ONUs),  so the power budget of 25G will not be a problem and it can bear lower ER.

 

Of course, the 25G ONUs may need to tolerate higher overload, but the downstream is continuous mode , the power level is constant (will only changes when you upgrade OLT), I think it can be overcome  without distinct extra cost.

 

 

发件人: frank effenberger [mailto:frank.effenberger@xxxxxxxxxx]
发送时间: 20171026 21:56
收件人: STDS-802-3-NGEPON@xxxxxxxxxxxxxxxxx
主题: Re: [802.3_NGEPON] Call for agenda items -- weekly consensus building for IEEE 802.3ca (100G-EPON)

 

John and All,

 

To my understanding, all the operators that have spoken on the topic have said that they will only upgrade when the new system provides at least 4x of the capacity of their current system.

IF THIS IS A TRUE STATEMENT (Operators, please confirm), then I can see two separate evolution paths:

 

1/2.5G-PON à 10G-PON à 50G-PON

or

1/2.5G-PON à 25G-PON à 100G-PON

 

In other words, either you will upgrade to 10G-PON, or 25G-PON, but not both.   

 

If we can make the further statement that we only need to support adjacent generation coexistence, then 50G needs to coexist with 10G, but NOT 25G.  Similarly, 25G needs to coexist with G-PON, but NOT 10G. 

 

Finally, it is attractive if we can use a single wavelength for 50G.  Technically, this seems quite possible – all we need to deal with is the power budget, but that is something that can be overcome with technology advance. 

 

This then suggests an obvious wavelength plan, where the 1270nm upstream would be used for 25G, and the 1310nm upstream would be used for 50G.  

 

You may fire when ready.  Let me be the first one to shoot:

A big open question here is the asymmetric system combinations.  The discussion above talks about “50G” as if it is one system, but it really isn’t: it includes both 50G symmetric and 50G/25G asymmetric. 

I need to hear from the operators on their opinions of the asymmetric system, and if that is of interest.  One possible relief of this would be to say: the asymmetric system would be supported using TDMA coexistence system.  That is, the 25/10 ONUs would use the same upstream wavelength as the 25/25G ONUs; and the 50/25 would use the same wavelength as the 50/50G ONUs. 

 

 

Sincerely,

Frank E.

 

 

From: John Johnson [mailto:000007ff7d378f43-dmarc-request@xxxxxxxx]
Sent: Thursday, October 26, 2017 9:30 AM
To: STDS-802-3-NGEPON@xxxxxxxxxxxxxxxxx
Subject: Re: [802.3_NGEPON] Call for agenda items -- weekly consensus building for IEEE 802.3ca (100G-EPON)

 

These presentations and previous discussions bring up the question of whether there will ever be any market for 50G-EPON OLTs, and if there is, whether it would need to coexist with legacy PONs.  China Telecom wants to jump to 100G OLTs on top of XG(S)-PON networks, so 100G-EPON needs to coexist with 10G PON.  Upgrade of GPON directly to 25G-EPON with WDM coexistence is a reasonable scenario, but it then seems likely that the next step in the far distant future would be 100G-EPON, presumably without a need for further GPON coexistence.  

 

So 100G-EPON needs WDM CE with 10G-EPON, 25G-EPON needs WDM CE with GPON, and it's not clear what CE is needed for 50G-EPON, or if there is any market for 50G-EPON at all.  If we can agree that 50G-EPON is likely not going to be deployed in any significant volume, then we might also agree that the CE requirements may be relaxed for 50G-EPON.  Comments?

 

Question for Dekun:  What is the view of China Telecom regarding TDMA of 25G and 50G on the same wavelength for US and DS?  Is this an acceptable solution from a capacity standpoint?  

 

One comment for dual bit rate downstream is that it is difficult to change OMA and ER on the fly without adding extra settling time between frames, so the NRZ signal may have to be limited to the lower ER that is used for PAM4 (~5dB).  

 

Regards,

John

 

On Thu, Oct 26, 2017 at 5:53 AM, Liudekun <liudekun@xxxxxxxxxx> wrote:

Curtis, all,

 

I will bring in a contribution on 25G/50G dual rate channels if the time slot is still possible

 

( I hope I dont have a problem to access the telephone meeting from the network in the other side of the pacific ocean)  

 

Best regards

Dekun Liu

____________________________________________________

Advanced Access Technologies Dept. 网络研究接入技术部

Huawei Technologies Co., Ltd. 华为技术有限公司Company_logo
  Phone: +86 027-59267217  Email: liudekun@xxxxxxxxxx

湖北武汉市关山一路光谷软件园A7-9 邮编:430074
Huawei Technologies Co., Ltd.

A7-9 Wuhan Optical Valley Software Park,Guan Shan Road,Wuhan,Hubei, P.R.China

 

 

From: Curtis Knittle [mailto:C.Knittle@xxxxxxxxxxxxx]
Sent: Monday, October 23, 2017 11:13 PM
To: STDS-802-3-NGEPON@xxxxxxxxxxxxxxxxx
Subject: [802.3_NGEPON] Call for agenda items -- weekly consensus building for IEEE 802.3ca (100G-EPON)

 

Dear Colleagues,

 

We have the 100G-EPON (IEEE 802.3ca) consensus building meeting scheduled for this Thursday, October 26, 11:30 am 1:00 pm MDT. Please let me know by 5:00 pm MDT Wednesday if you have any topics for the agenda.

 

As a reminder from the closing report of our last F2F meeting in Charlotte, the following topics were deemed important for contributions:

 

  • Actions:
    • Power budget
      • Close on 25Gbps power budget (Ed/Dekun/Umeda)
      • Characterization method for both transmitter and receiver (Dekun, John J, Vincent, Junwen, Frank, Greg, Umeda-san)
    • MPCP/MPRS
      • Discovery process / configuration / ONU capabilities field in RegREQ (Glen Oct 12 Glen to doodle poll)
    • Fault Tolerance/Recovery/Loss of Channel
    • Wavelength plan
      • Peacemaker plan (Glen/Frank/Ed/Yong)
    • SOA preamp/dynamic range
      • Cost and feasibility still needs to be demonstrated (Dekun)
    • Line coding (Postponed until after FEC decision)
    • FEC information
      • Comparisons of parity matrices and different FEC (Mark/Bill/Bo)
      • Loud/soft ONU analysis (Bill)
      • Operating in burst mode at 1E-2 (Vincent)
      • Latency (Shan)
      • Define a common set of assumptions and parameters common to all FEC analyses
        • Mark to send email after noodling
    • SERDES/CDR
      • Check margin for 10^-2 input BER (Yin, Ryan)
    • 50G single wavelength
      • Technical feasibility of upstream bursts (Critical path item)
      • 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

c.knittle@xxxxxxxxxxxxx

 

Stay up to date with CableLabs: Read the blog and follow us on Twitter

 

 

JPEG image