3. Channel Codingand ModulationCurrentandfuturekeyrequirements Today Mostly 10G OOK 40G was a transition step to coherent, DSP-based technologies OOK, DPSK, DQPSK, PM-QPSK Commercial success and further lifetime questionable ! PM-QPSK 100G coherent (1st generation) picking up What‘s next ? 400G, 1T ? Maximise spectral efficiency vs. reach ? Minimise costs ! Get flexibility by Software Defined Optics (SDO) Today RS-FEC Concat.-FEC Turbo S-FEC channelcoding andmodulation ? OOK DB DPSK PM-QPSK 2.5G 10G 40G 100G
4. Channel Codingand ModulationRelative COGS oftransponders 100G coh normalized on 2011 10G cost needstobe adressed ! Cost efficiency of 40G questionable – need low cost 100G option !
5. Coding Add. NCG smaller vs. increasing OH 1 dB add. by soft-in decoding Modulation single pol., SSMF, 100km spans, ideal Raman, no DC, WDM with 5 channels Channel Codingand ModulationWherearethelimits ? 14 9 8 13 7 12 6 11 5 10 4 9 3 8 2 7 1 6 0 5 0 5 10 15 1 1.1 1.2 1.3 Shannon limit Gaussianch. fibercapacitylimit [1] 500km Shannon limit for ideal FEC 256QAM Shannon limit soft 2000km 2 bit hard 64QAM 8000km Spectral Efficiency (bits/s/Hz) Net CodingGain [dB] for BER=1e-15 16QAM 100G implementations 8PSK QPSK Shannon limit Gaussianchannel G.709 BPSK -1.5 20 25 1.4 1.5 transmission rate SNR/bit (dB) [1] Essiambre, et al., “Capacity Limits of Optical Fiber Networks,” JLT, vol. 28, no. 4, Feb. 2010. Scale by Superchannel/OFDM & spatial diversity (polarization/fiber)
6. Channel Coding and ModulationWhat do we need to get there ? High speed DSPs/DACs/ADCs : power limitation ! Photonic Integration Photonics are dominating optical transceiver size & cost Options : InP, hybrid, CMOS photonics Adapted from Fujitsu Microelectronics ≈1 mm Oclaro : 40 Gb/s InP DQPSK Encoding Chip
7. Channel Coding and Modulation400G ? 480 Gb/s (incl. 15% FEC OH) Nyquist WDM spectral shaping Total BW = #subcarriers x symbol rate Only noise limitations considered Overall power remains constant Channel granularity: 50 GHz PM-64QAM Capacity x reach = const. PM-8QAM PM-QPSK PM-16QAM PS-QPSK 100GPM-QPSK
8. Channel Coding and Modulation1T ? 1200 Gb/s (incl. 15% FEC OH) Nyquist WDM spectral shaping Total BW = #subcarriers x symbol rate Only noise limitations considered Overall power remains constant Channel granularity: 50 GHz Capacity x reach = const. PM-16QAM PM-8QAM PS-16QAM PM-QPSK 100GPM-QPSK
10. Channel Codingand ModulationµWave Radio – Adaptive Modulation & Coding (AMC) AMC to offer variable link ranges, data rates, availability @ BER 1E-11 All overhead considered 16-QAM, 25 min non-availability/year VBR CBR 64-QAM, 115 min non-availability/year UBR VBR CBR 4-QAM, 5 min non-availability/year CBR Hitless switching Between PHY modes FIXED sliced spectrum given Source : Marconi (now Ericsson) Hitless AMC for flexible usage of a FIXED sliced spectrum
15. Optical Layer – Line SystemFully integrated EDFA/Raman amplification Performance of different hybrid amplifiers Improved net noise figures by hybrid amplification
16. Gain controlled Output power=+21dBm & NF=4.5dB Transient event =1usec & Add/Drop=16dB Gain excursion<1.5dB Self saturated Output power=+21dBm & NF=4.5dB Transient event =1usec & Add/Drop=19dB Gain excursion<0.4dB Optical Layer – Line SystemTransient suppression Increased transient suppression by fill lasers or self-saturation
17. Optical Layer – Line SystemSummary Flexibility, enhanced system margin supported by Reduced losses ROADM design, low loss fiber ? Improved OSNR Hybrid amplification Increase transient suppression Self-saturated EDFAs Fast VOA integrated with EDFAs Component needs : High power pump sources Low relative intensity noise Raman pump sources Today optical layer gain/power control variable gain control Transient immune, hybrid amplification line system 8 ch 96 ch non-DCx 160 ch C+L EDFA Raman hybrid
27. … … Optical Layer - SwitchingROADMs … andthisishowitcouldlooklike IL = 9 dB Per degree No single-point-of-failure Scalable in directions and A/D capacity Minimum loss IL = 9 dB … … Here : Twin WSS architecture Could be splitter (check IL and Isolation) 1x16 WSS 1x16 WSS Line 9 ports Up to 96 channels per port … but : all WSS need to beFlexgrid and are not available today IL = 6 dB A/D 1x4 Comb 1x4 Comb … … Scaling to reach full add/drop capacity w/o only 25% A/D capacity (need 768:24 = 32 feeds !) passive fiber arrangement IL = 1 dB … IL = 9 dB 8 x24 WSS 8 x24 WSS 100% add/drop capacity for all degrees (768 ch.) …
28. Optical Layer - SwitchingROADMs … it‘s all aboutcompromises ! IL = 6 dB Restrict to max. 6 degrees … … or scale with couplers On line side or WSS output side Insertion Loss !!! Per degree IL = 6 dB … … … … 1x9 WSS 1x9 WSS Line 4 ports Upto 96 channels per port Many different options (incl. reduction of A/D capacity) Cascading WSSs Combining WSS and multicast switches (PLC) Monolithic switch plus splitter and filters … Insertion Loss : in any case multiple amplifiers included ! A/D
30. Optical Layer – SwitchingSummary CDCF ROADMs are here today ! Ideal components not available today Realization with supporting technologies possible Avoid internal amplification as much as possible Ensure steep passbands, proper isolation Component needs : Line side WSS : 1xN Flexgrid with N as large as possible A/D WSS : NxM with M as large as possible Optical Power Monitoring Must be Flexgrid too Needed on line and add/drop sites Today opticallayer 100 Ghz Flexgrid 50 Ghz NG-CDCF switching colorless contenionless directionless FOADM 2D-ROADM MD-ROADM
31. Protocols and Multi-Layer IntegrationCurrent and future key requirements G.709 / OTN Scalable wrapping, multiplexing and switching technology Evolved to be more Ethernet friendly ODUflex support channelization of TDM & packet interfaces Hitless resizing provides for in-service channel sizing Need to support future bitrates and transparent timing Ethernet, MPLS-TP, MPLS All evolving and having their play Multi-layer integration is the key challenge Today T-MPLS MPLS-TP ? transport packet EFM CFM Y.1731 1G 10G 40G/100G Protocols ? SONET SDH TDM G.709v3 G.709v1 G.709v2
33. Protocols and Multi-Layer IntegrationMPLS-TP and Ethernet Both, Ethernet and MPLS extended with Transport Profiles (TP) OAM, protection, traffic engineering, static and dynamic options, … Comparison is difficult MPLS-TP might have benefits in MPLS interworking (but …) Ethernet is the data link layer, always ! The clever bit is to ensure seamless interworking MPLS, VPLS Service VLAN MPLS PW Tunnel VLAN Link VLAN MPLS Link Ethernet, GFP ODU switching OTN Framing, FEC, OAM Optical switching and transport Multiple options to achieve the same !
34. Protocolsand multi-layerintegrationMulti-layer network study - results US, 46 Nodes, 18 Tb/s, 1:1 packet:TDM-> 2:1 10GbE (grey) … OTU2 (grey) typicalrange 23%savings … Packet Switch(MPLS) OTU2 (grey) … 10GbE (grey) … 10GbE (grey) … OTU2 (grey) … Hybrid Packet/ Circuit Switch (MPLS/ODU) … OTU4 (grey) Packet Switch(MPLS) Circuit Switch (ODU) Circuit Switch (ODU) OTU4 (colored) … … 96 l DWDM 96 l DWDM Contentionless MD-ROADM OTU4 (colored) OTU4 (colored) … … … … 96 l DWDM 96 l DWDM 96 l DWDM 96 l DWDM Contentionless MD-ROADM Contentionless MD-ROADM Up to 23% savings with an integrated switch Autenrieth, et.al., “Benefits of Integrated Packet/Circuit/Wavelength Switches in Next-Generation Optical Core Networks”, NFOEC 2011, NMC4
37. Management andControlKey enabler : multi-x control plane Multi-Degree Auto-discovery of topology (OSPF-TE) Constraint-aware path computation Automated signaling (RSVP-TE) Mesh networking, agile endpoint selection, tunable origination and regeneration Multi-Region Transport networks growing in size and complexity Formerly islands, regional networks are linking up Multi-Layer Flexible, agile WDM transport layer, integrated Ethernet/MPLS layer, integrated OTN TDM layer Multi-Service Automated Restoration Fault detection/reporting, dynamic channel re-route Embedded Intelligence in Every Element Multi-Vendor Protocol standardization, proven Interoperability GMPLS core, OIF & ASON compatibility
38. Management and ControlOne Tool to handle the complexity : PCE Architecture Separate where computation is needed from where it’s performed Path Computation Client (PCC) Requesting path computation services (can be NE, NMS, Tool, PCE) Path Computation Element (PCE) Performs path computations on behalf of PCCs or other PCEs Standardized toolbox approach Distributed, centralized, hybrid approaches Sees nodes <E,F,G,H> Sees nodes <A,B,C,D,E> “compute A to H” “compute E to H” Sees self PCE PCE PCC “E->F->G->H” “A->B->C->D -> E->F->G->H” Addressing the complexity in a standardized way
39. Management andControlSummary Interoperable network automation by standardized architecture IETF: Routing Area, multiple working groups e.g. PCE OIF: User-to-Network / Network-to-Network IAs (UNI/E-NNI) ITU-T: Automatically Switched Optical Network (ASON) TMF : Management frameworks and interfaces (e.g. MTOSI) Future needs are endless ! Multi-layer definitions/interactions resource sharing, provisioning, protection, restoration, OAM interaction, … OTN extensions Optical constraints (wavelength, path, OSNR,…) … many more ! Today OSS Integration Automated top down multi-layercontrol management andcontrol Corba TL-1 XML/MTOSI SNMP Q ASON GMPLS
40. SummaryThe programmable & automated optical network Today OSS Integration Automated top down multi-layer control management and control Corba TL-1 XML/MTOSI SNMP Q ASON GMPLS T-MPLS MPLS-TP MPLS/MPLS-TP transport packet EFM CFM Y.1731 IEEE 802.1/2/3 1G 10G 40G/100G 400G/1T Protocols It won’t get boring ! SONET SDH 400G/1T ? TDM G.709v3 G.709v1 G.709v2 100 Ghz Flexgrid 50 Ghz NG-CDCF switching colorless contenionless directionless FOADM 2D-ROADM MD-ROADM optical layer gain/power control variable gain control Transient imune, hybrid amplification line system 8 ch 96 ch non-DCx 160 ch C+L EDFA Raman hybrid Integrate RS-FEC Concat.-FEC Turbo S-FEC channel coding and modulation SDO OOK DB DPSK PM-QPSK 2.5G 10G 40G 100G