Скачать презентацию The Path to 100 G New challenges for Скачать презентацию The Path to 100 G New challenges for

144c6a64285a100904e1277559619e5a.ppt

  • Количество слайдов: 20

The Path to 100 G New challenges for 40 G and 100 G Networks The Path to 100 G New challenges for 40 G and 100 G Networks Braodband Technology Event Arthur Moll Rotterdam, November 10 th, 2009 BDM T&D EMEA

Why is 40 G - 100 G introduction so difficult? § We are reaching Why is 40 G - 100 G introduction so difficult? § We are reaching the limits of the physics § Limits of the optical fiber § Limits of the optics § Traditional technologies (NRZ, DWDM, EDFAs, § § ROADMs, etc) have issues at 100 G Old Rule of Thumb 4 x. Data Rate = 2, 5 x. Costs § Still valid for. STM-16 to STM-64 § Not valid STM-64 to STM-256 (4 x. Data = 5 x Costs) Mapping of 1 Gig. E and 10 Gig. E is difficult § E. g. 10 Gig. E with GFP in OTU 2 muxed in OTU 3 2

From 10 G to 40 G or 100 G – Fiber Limit T STM-64 From 10 G to 40 G or 100 G – Fiber Limit T STM-64 STM-256 OTU-4 § § T= 100 ps T= 25 ps T= 8 ps CD Limit: 16 x less (120 x) PMD Limit: 4 x less (11 x) 2 nd order PMD can not be neglected anymore Makes some current network structure un-usable 3

From 10 G to 40 G or 100 G – Optic Limit 10 Gbps From 10 G to 40 G or 100 G – Optic Limit 10 Gbps 40 Gbps 100 Gbps? ? ? § § § Faster Modulation = Broader Spectrum - FWHM Spec meaningless! Broader Spectrum ->Peak Power is not the answer -> Integrated Power OSNR needs to be defined&measured differently Curve shape needs to be considered. Can we still do WDM? Risk of Cross-Talk. 4

“Standard” OSNR measurement method IEC 61280 -2 -9 FLAT Noise floor 5 “Standard” OSNR measurement method IEC 61280 -2 -9 FLAT Noise floor 5

 • OSNR measurement method Special case of filtered signals • Real noise • • OSNR measurement method Special case of filtered signals • Real noise • Interpolated noise • For filtered signals (ROADMs), Interpolation method under-estimates the noise level as noise is « carved » with the signal by the filter • Different paths have different Noise Contribution 6

EXFO Innovative In-band OSNR method Polarization diversity method • Theory: Signal is polarized and EXFO Innovative In-band OSNR method Polarization diversity method • Theory: Signal is polarized and noise is depolarized • EXFO OSA have a polarization controller and polarization beam splitter at the input , allowing automated In-band OSNR measurement • Power vs polarization • Ppeak • Signal • OSNR • Noise • In-band • PNoise 7

What are conditions for usable 100 G System? Conditions for usable 100 G Line What are conditions for usable 100 G System? Conditions for usable 100 G Line Side Transmision-System: § More tolerant against CD&PMD § Fit into WDM Grid – 100 GHz, but 50 GHz is better § Co-operate with existing 10 G/40 G channels § Compatible with EDFAs and ROADMs § High OSNR § § High Rx Sensitivity Low EDFA Noise The answer: New Transmission System with new Modulation format 8

NRZ/RZ Intensity modulation NRZ-OOK (On/Off Keying) vs. RZ-OOK (On/Off Keying) NRZ-OOK Rx The intensity NRZ/RZ Intensity modulation NRZ-OOK (On/Off Keying) vs. RZ-OOK (On/Off Keying) NRZ-OOK Rx The intensity modulation is easily detected by direct detection with a photo detector 9

Phase Modulation: DPSK (Differential Phase Shift Keying): 0 p 0 0 0 p 0 Phase Modulation: DPSK (Differential Phase Shift Keying): 0 p 0 0 0 p 0 1 1 0 0 1 A logical « 1 » and a « 0 » have light A « 1 » is represented by a phase shift by p. 10

Phase Modulation: QPSK-DQPSK/DQPSK modulation QPSK DQPSK 4 phases 2 bits/symbol 11 Phase Modulation: QPSK-DQPSK/DQPSK modulation QPSK DQPSK 4 phases 2 bits/symbol 11

Pol. Muxed – Phase Modulation: DPQPSK DP-QPSK Modulation Two independent Data Streams Multiplexed with Pol. Muxed – Phase Modulation: DPQPSK DP-QPSK Modulation Two independent Data Streams Multiplexed with Polarization To achieve 112 Gbit/s typical with 28 Gbaud PM-(D)QPSK Very complex, cost intensive, but really robust! 12

Performance Overview Ref: Can 100 Gb/s wavelengths be deployed using 10 Gb/s engineering rules? Performance Overview Ref: Can 100 Gb/s wavelengths be deployed using 10 Gb/s engineering rules? Strata. Light Communications Inc, Cisco Systems Inc 13

DP-QPSK Testing – Constellation Analyser Confidential 14 DP-QPSK Testing – Constellation Analyser Confidential 14

IEEE 802. 3 ba Pluggable Modules § CFP form factor package (86 x 127 IEEE 802. 3 ba Pluggable Modules § CFP form factor package (86 x 127 x 14 mm / 3. 4”x 5. 0”x 0. 55”) § § 100 Gb. E, 40 km on SMF (4 x 25 G LAN WDM, centered at 1305 nm) 100 Gb. E, 10 km on SMF (4 x 25 G LAN WDM, centered at 1305 nm) 40 Gb. E, 10 km on SMF (4 x 10 G CWDM, centered at 1305 nm) 100 Gb. E, 10 km on SMF (10 x 10 G CWDM, centered at 1550 nm) § § § 100 Gb. E, 100 m on MMF (850 nm parallel optics, 10 x 10 G) CXP form factor (approx 20 x 54 x 11 mm / 0. 78”x 2. 13”x 0. 43”) § § § From Santur Corporation 100 Gb. E, 100 m on MMF (850 nm parallel optics, 10 x 10 G) 100 Gb. E, 10 m on active cable QSFP form factor (18. 4 x 72 x 8. 5 mm / 0. 72”x 2. 8”x 0. 33”) § § 40 Gb. E, 100 m on MMF (850 nm parallel optics, 4 x 10 G) 40 Gb. E, 10 m on active cable CONFIDENTIAL 15

MAC & Mux/Demux in PCS Lanes PHY 12 11 10 9 8 7 6 MAC & Mux/Demux in PCS Lanes PHY 12 11 10 9 8 7 6 5 4 3 2 1 0 19 18 17 aka: CFP 16 15 14 100 Gb. E serial bit stream b bits 12 8 4 13 Mux/Demux 11 0 6 14 Mux/Demux (2: 5 5: 10) 17 2 3 19 15 5 1 10 7 16 9 18 13 Mux/Demux (2: 5 5: 10) 8 4 0 16 12 3 19 15 11 7 10 6 2 18 14 10 6 2 18 14 12 8 4 # 0 0 16 # 1 12 3 # 2 19 15 # 3 11 7 # 4 # 5 # 6 13 # 7 5 1 17 9 13 9 5 1 17 # 8 # 9 Each PCS Lane is 5 G bit stream 16

Implications? What needs to be Skew= tested? Propagation Difference, § 148 Mio packets/s or Implications? What needs to be Skew= tested? Propagation Difference, § 148 Mio packets/s or 3 x. DVDs per sec § § BERT of complete link BERT per PCS Lane Bit delay, Offset § PCS Lane concept is complex and has trouble zone § § § PCS Lane Marker (Order, Mapping) Lane Skew Mapping 10 x 10 G elec. into 4 x 25 G bring new challenges § Ethernet Parameter (standard) § § Frame Size distribution, Errors (FCS, Runt) Thru. Put, Frame Loss 17

FTB-85100 G – First portable 100 G Tester BERT Ethernet & IP Testing § FTB-85100 G – First portable 100 G Tester BERT Ethernet & IP Testing § § On physical lane, PCS Lane or trunk PCS Error injection and monitoring § § Packet Definition § § § User defined rate utilization User defined MAC and IP address User defined packet size up to 16 k § Jabber, giant, runt, oversize, FCS, total frames Bit error in Physical Lane or in PCS Lane Error injection and monitor Invalid 64 b/66 b or unsupported Layer 2 PCS code § Payload and FCS Invalid or duplicate PCS Lane markers Ethernet statistics PCS Lanes § Layer 1 error/alarm injection and monitoring Monitors lane marker period and reports any changes to PCS lane table User defined or random mapping supported IP statistics § Multicast, broadcast, unicast, total Actual bandwidth usage CONFIDENTIAL statistics 18

85100 G – Interface Specifications Client Interfaces § § Very high CFP connector insertion 85100 G – Interface Specifications Client Interfaces § § Very high CFP connector insertion rating 100 G CFP, 100 G CXP, 40 G CFP, 40 G QSFP Clocking § § Stratum-3 e clock with +/- 115 ppm user controllable offset Support 100 G /112 G and 40 G/43 G rates OTU 3 and OTU 4 hardware ready § § Unframed BERT (PRBS) at OTN rate Framed OTN with parallel optics support on roadmap Pluggable interface controls & monitoring § § § Detailed CFP interface control Control: laser on/off & amplitude (per lane) Monitoring: Input level (per lane), Rx frequency CONFIDENTIAL 19

Questions? Questions?