Reconsiderations on 50G and 100G EPON IEEE Berlin meeting

Transcription

1 Reconsiderations on and 100G EPON 2017 IEEE Berlin meeting 1

2 Contributors China Telecom: Wang Bo, Zhang dezhi China Unicom: Zhou xiaoxia, Shen shikui, Guo lin, Shao Yan China Mobile: Zhang dechao, Wang Lei CAICT: Cheng qiang CTTL: liu Deqiang Nokia Shanghai Bell: Chen xiao ZTE: Yuan liquan Fiberhome: Wang suyi, Huang Yuanbo Huawei: Liu dekun, Lin Huafeng Source Photonics : Jiang Xu, Shuai Xin, Guo Songtao Hisense: Zhang Hua Accelink: Chen Xuguang Inno Light: Song Yan 2017 IEEE Berlin meeting 2

3 PON evolution in China Market Next generation PON after 10Gb/s is expected to satisfy the service requirements around 2025 From massive deployment point of view, most probably to evolve from 10G platform (OLT) directly to /100G TDM PON (EPON and GPON) has been successfully deployed and operated in China for many years mainly for residential broadband access, showing great advantages. Hence we suggest to consider and evaluate the possibility of a single wavelength system with enhanced line rate such as b/s, when developing next generation PON system 2017 IEEE Berlin meeting 3

4 System Architectures for comparison PON(2) PON(1) PON 25G Tx1 25G Tx2 25G Rx1 25G Rx2 λ1, λ2 2*25G splitter Rx1 Tx1 Rx2 Tx2 Rx1 Tx1 Rx2 Tx2 ONU ONU VS PON OLT TDM-PON λ0 b/s splitter Rx Tx Rx Tx Rx Tx Rx Tx ONU ONU ONU ONU PON may have two architecture options: TDM-PON (named as PON(1) for communication convenience) 2x25G based (named as PON(2) for communication convenience) Correspondingly, 100G PON may have also two architecture options: 4x25G based 100G PON (named as 100G PON(4) for communication convenience) 2x based (named as 100G PON(2) for communication convenience) 2017 IEEE Berlin meeting 4

5 Candidate technologies for bps Tx/Rx(1) PAM-4 Reuse the industry chain in 200GbE and 400GbE MSB {0,1} LSB {0,1} 25G logic 4-level encoder 25G APD BMR 25G EML Tx* {0,1,2,3} Diplexer EDB Reuse existing 25G optics in data center logic 25G Tx* 25G APD Rx* Diplexer There are several candidate technologies for b/s per wavelength, but utilizing the existing 25G optics assisted with advanced modulation seems the most promising technologies 2017 IEEE Berlin meeting 5

6 Candidate technologies for bps Tx/Rx(2) Discrete Multitone Modulation using 10G optics bit loading on subcarriers logic +DSP 10G APD BMR 10G EML Tx* Diplexer 0 10 GHz QPSK and QAM bit loading of 10G spectrum using DSP at Tx and Rx Receiver sensitivity b2b at BER=10-3 : Gbps =-21.5 dbm; Gbps =-15.5 dbm 2017 IEEE Berlin meeting 6

7 PON Wavelength plan comparison From wavelength aspect, upstream of PON is the emphasis in order to insure low cost ONU. Downstream wavelength of PON is expected to be cooled and narrow band, and is much easier to get appropriate wavelength resource. 25Gb/s per lane b/s per lane UP DN UP DN PON(2) PON(1) (Note: the value of the wavelengths are just examples) Compared with PON(2), wavelength demand and constraints of PON(1) are less. And it will reduce complexity of wavelength allocation, and may loosen the channel width of upstream (even uncooled is possible) IEEE Berlin meeting 7

8 Comparison of PON(1) and PON(2) Take b/s PAM4(25GBaud) as an example (Re-using 200GE and 400GE industry chain) 2*25G 1* 25G Tx1 25G Tx2 25G Rx1 25G Rx2 VS logic 25G Tx 25G APD Rx Diplexer 25G LD 25G APD cooled 2~3nm channel width 20nm channel width uncooled 2017 IEEE Berlin meeting 8

9 mux λ demux PON(2) PON(1) PON(1) vs PON(2) Power budget analysis λ1 25G APD λ2 25G APD λ 1 25G Tx λ 2 25GTx OA OA λ demux -1dB mux +8.8dBm 25G APD 25G LD SOA -1dB SOA TFF diplexer TFF diplexe r +7.3dBm -0.5dB 29dB +1.5dB TDP -1dB -23.2dBm@1e-3 TFF diplexer -0.5dB 2dB harder than 25G PON 29dB +1.5dB TDP TFF diplexer -1dB -20.7dBm@1e dBm OA -0.5dB -0.5dB +10.3dBm 5~6 db harder than 25G PON (NRZ) 2017 IEEE Berlin meeting dBm@1e-3 λ1 25G APD λ2 25G APD λ 1 25G Tx 25G APD λ 2 25GTx -21.2dBm@1e-2 25G LD PON(1) based on 25G optics with PAM4 modulation is also possible to meet 29dB power budget with optical amplifier and enhanced FEC. NRZ modulation PAM4 modulation T 0 /3 T 0 5dB sensitivity loss compared with NRZ in theory T 0

10 Industry Chain in Ethernet PtP Network 10GbE 25GbE 40GbE be 100GbE 200GbE 400GbE 500m 10G G X - 100G X100G - 4X100G 2km 10G G G - - 1X40G X - 4X 8X 100G km 10G 1X10G - 4X10G G - 1X25G - - 4X25G X - 4X 8X 100G km 10G 1X10G - 4X10G G - 1X25G - - 4X25G G More and more Ethernet work are based on b/s per lane 2017 IEEE Berlin meeting 10

11 100G PON(2) vs 100G PON(4) comparison 25Gb/s per lane b/s per lane 100G UP 100G DN 100G UP 100G DN 100G PON(4) 100G PON(2) * 25G 2* 4 2 transmitter 25GEML/DML 25G EML/DML Receiver 4 25G APD 2 25G APD Cool or uncool Cooled Uncooled Mux/Demux 4 cascaded filters 2 cascaded filters Modulation Format NRZ PAM4 PAM encoder and Electronics 25G LDD, TIA Decoder Extra required launch Power compared 4dB 7~8dB with 25G PON Reusing existing Industry Chain 25GE & 100GE E,200GE, 400GE There is no clear advantage for 100G PON(4) compared with 100G PON(2), it deserves more study on b/s single wavelength 2017 IEEE Berlin meeting 11

12 Motions Motion 1 The task force should analyze and compare the following solutions for PON: 1) Single wavelength TDM-PON with b/s line rate 2) Two-wavelength TDM/WDM-PON with 25Gb/s line rate per lane and choose the best one for EPON. Motion 2 The task force should analyze and compare the following solutions for 100G PON: 1) Two wavelength TDM/WDM-PON with b/s line rate per lane 2) Four wavelength TDM/WDM-PON with 25Gb/s line rate per lane and choose the best one for 100G EPON IEEE Berlin meeting 12

13 Backup-PAM4 feasibility reference Optical transmission feasibility for 400GbE extended reach PMD IEEE Berlin meeting 13