State and future of optical transport networks Ro

State and future of optical transport networks Ro. Edu. Net conference Cluj, 28 th August 2008 Andreas Hegers Director Solutions Architecture and Strategy Metro Ethernet Networks 1

Agenda • Market trends • Main Services and Bandwidth drivers • Technology trends • Key network requirements • Ways to build a future-proof transmission network • DWDM transmission • 10 G - Still the baseline • 40 G - The next big thing • 100 G - On the horizon • Networking flexibility • • ROADMs - Photonic flexibility OTN - The successor of SDH (? ) L 2 - Embedded data capabilities Control plane - Gluing it all together • Network example Ro. Edu. Net • Outlook and summary 2

Agenda • Market trends • Main Services and Bandwidth drivers • Technology trends • Key network requirements • Ways to build a future-proof transmission network • DWDM transmission • 10 G - Still the baseline • 40 G - The next big thing • 100 G - On the horizon • Networking flexibility • • ROADMs - Photonic flexibility OTN - The successor of SDH (? ) L 2 - Embedded data capabilities Control plane - Gluing it all together • Network example Ro. Edu. Net • Outlook and summary 3

The Need for Speed You. Tube today uses as much bandwidth as the entire Internet in 2000: • 200 Tbytes of traffic daily By 2010, 27% of Business access will require 100 M 10 G Ethernet EADs (Millions) 3. 0 2. 5 2. 0 1. 5 More than 70% of U. S. Internet users, streamed or downloaded Web video in 2007. Source: Infonetics 2 H 2007 10/100 M 44% CAGR (07 -10) 10/100 M 1 G 36% CAGR (07 -10) 1 G 10 G 415% CAGR (07 -10) 10 G 1. 0 Storage bandwidth growth: • 6, 000 Terabytes 2008 -> > 16, 000 Terabytes 2010 0. 5 0. 0 2007 2008 2009 2010 Gartner Group, Oct 2006 We are in the middle of massive growth of networks where bandwidth requirements are exploding. 4 Source: Infonetics & Nortel Analysis

100 Gbps Terrestrial Gbps Routes 21. 4% 10. 5% 5. 4% • Sum of capacities from various user groups builds need for 40 Gb/s and eventually 100 Gb/s links 4. 4% 40 Gbps • Video and voice services drive more stringent Qo. S expectations 10 Gbps 20 09 20 08 20 07 20 06 20 05 20 04 20 03 20 02 20 01 2. 5 Gbps 20 00 Gbps Routes in Millions (represents new ports shipped) Impact on Transport Source: Infonetics, 2 Q 06 & Nortel internal 100 G study 5 • Optical (data center) services require multi-Gb/s over full time or on-demand connections

Key Enabling Technologies – Optical Access Metro Transport & Service Management Business Services Access Ethernet PDH / SDH IP L H Core Optical Modem: Fully Tuneable 10 40 100 Gbs Adaptive Distortion Mitigation (CD, PMD/PDL, Non-linearities) Residential Services Access Agile Packet Optical WSS based ROADM for network agility, multi-way branching Wireless Backhaul Gb. E Storage Leased Lines BB Business Services Access 6 Converged L 0/L 1/L 2 in single platform for fully flexible capacity allocation Photonic domain control intelligence for fully automated line control and simplified end to end provisioning

Key Enabling Technologies – Ethernet Access Metro Transport & Service Management Business Services Access Ethernet PDH / SDH IP Residential Services Access L H Core IP / MPLS PBB-TE (PBB-TE) Deterministic Ethernet Circuits GMPLS provisioning efficiencies Carrier Ethernet 802. 1 ag and Y 1731 carrier grade operations and instrumentation Agile Optical Wireless Backhaul Gb. E Storage Leased Lines BB Business Services Access 7 PBB – secure, scalable clear demarcation between customer and provider addressing Ethernet with the Efficiencies of Packet and the Robustness of SDH

Agenda • Market trends • Main Services and Bandwidth drivers • Technology trends • Key network requirements • Ways to build a future-proof transmission network • DWDM transmission • 10 G - Still the baseline • 40 G - The next big thing • 100 G - On the horizon • Networking flexibility • • ROADMs - Photonic flexibility OTN - The successor of SDH (? ) L 2 - Embedded data capabilities Control plane - Gluing it all together • Network example Ro. Edu. Net • Outlook and summary 8

Network Simplification Through Innovation DSCM C-Band DSCM OEO L-Band … OEO OEO DSCM AMP Node DSCM OEO DSCM … Terminal Node DSCM … … OEO OEO OADM Node Electrical Signal Processing Advanced FEC Advanced line and modem Terminal Node OEO 9 OEO … … OEO AMP Node Remove/Minimize DSCMs & Amps, Increase PMD Tolerance, Eliminate complex engineering rules (esp. OADM) Improved coding-gain Simple deployment & reconfiguration, reduced inventory & truck rolls e. ROADM Node OEO

Advanced E/O Modem Introduction 10 G e. DCO • Electrical Tx based dispersion compensation Wraptor FEC • Better than +/- 50, 000 ps/nm • 9. 4 d. B of coding gain • Real-time Fully automatic • 3 d. B > RS-8 • Raman avoidance 2003 2005 2008 Future Embedding Transmission Complexity into Electronics 10 10

Optical Pulse Transmission with Electronic Dispersion Compensation (e. DCO) on a 10 Gb/s link Conventional Optical Link with DCMs Rx Tx 1 span DCF DCF DCF = Dispersion Compensating Fiber, packaged as a DCM Nortel’s Next Generation Optical Link with CPL and e. DCO Tx Pre-Distorted, Eye Diagram 11 Rx Focused Eye Diagram (Zero Net Dispersion)

e. DCO Dispersion Scan 20 spans - 1. 600 km Dispersion [ps/nm] 27, 000 26, 750 26, 500 26, 250 26, 000 25, 750 25, 500 25, 250 25, 000 24, 750 24, 500 24, 250 24, 000 23, 750 23, 500 23, 250 23, 000 22, 750 22, 500 22, 250 22, 000 12 But what about 40 G…?

Fiber parameters - Things to know • The key fiber parameters to pay attention to are 1. Attenuation: 2. Chromatic Dispersion (CD): 3. Polarisation Mode Dispersion (PMD): For 40 G, the limiting factor is mostly PMD • Many carriers don‘t know the PMD values of their fiber, thus we have to stress the importance • The older a fiber, the higher usually it‘s PMD. One bad part will spoil the complete link • At 100 G, the situation is much worse for all 3, so a future proof solution is key 13

40 Gbps TDM – Challenges vs. 10 Gbps • 4 times the baud rate of 10 G TDM • Bit interval reduced from 100 ps to 25 ps • Circuit implementation significantly more challenging also need more complex materials • 4 times less light entering the receiver • 6 d. B drop in noise margin, may need RAMAN amplifiers • Increase optical spectrum occupied by a factor of 4 (to ~ 6 RZ) • Increased system impact of optical filters (OADM/ROADM) • 16 times less tolerant to chromatic dispersion • More stringent dispersion map • Increased installation difficulties, needs to be engineered day one • May need active CD compensators • 4 times less tolerance to PMD • May need PMD compensators • May need to select/match fiber based upon vintage, installation, etc… 40 G/ transmission has Significant Optical Challenges 14

Advanced E/O Modem Introduction e. DC 40 10 G e. DCO • Electrical Tx based dispersion compensation Wraptor FEC • Better than +/- 50, 000 ps/nm • 9. 4 d. B of coding gain • Real-time Fully automatic • 3 d. B > RS-8 • 2 -Pol QPSK 40 G • 10 Gbaud operation • +/- 50, 000 ps of CD compensation • Electrical PMD mitigation • 50 GHz OADM compatible • Raman avoidance 2003 15 15 2008 Future Embedding Transmission Complexity into Electronics

40 Gbps Dual Polarization QPSK • 40 Gbit/s on a single wavelength at 10 GBaud • Using Quadrature Phase Shift Keying (QPSK) • 2 bits/symbol: X 2 • 2 QPSK signals, one per polarization • 2 orthogonal polarizations: X 2 • World’s first fully integrated 40 G coherent digital receiver Dual Polarization Vertical Polarization Horizontal Polarization Dual Polarization • Propagates like a 10 Gbps signal • For non-linear impairments, dispersion tolerance, PMD tolerance, etc… QPSK X - polarization (0, 1) • Uses 10 G components: cost optimized, mature technologies with numerous vendors • Fully leverages existing 10 G Line Infrastructure • Same Reach – No RAMAN or reduction to overcome increase in noise I (1, 0) (0, 1) • No Dispersion Compensation required 16 (1, 1) QPSK Y- polarization • Same tolerance of cascaded ROADMs • Better PMD Performance than 10 G systems • All fiber that could be used for 10 G can now be used for 40 G Q (0, 0) I (1, 0) Rx Data Before DSP (1, 1) Rx Data After DSP

40 Gbps Dual Polarization QPSK 40 G Dual Polarization QPSK 50 GHz 10 G Conventional TDM 40 G Conventional TDM Frequency 40 G TDM Severely Impacted by Cascaded ROADMs System Severely Limited at 50 GHz-Spacing Carries Less Traffic 17

40 G Modulation Schemes Performance Comparison 10 G CD Tolerance [p. Sec/nm] DPSK CS-RZ DQPS K 2 -POL QPSK 1 Normalized Reach PSBT. 4 . 8 . 55 . 65 1 +/-400 PMD Tolerance [p. Sec] +/- 400 +/- 50, 000 3. 5 8 25 50 GHz 12 3 3 N/A 8 >23 100 GHz Filter/OADM Tolerance [# of ADM traversed] 15 12 8 8 8 >12 >23 But what about 100 G…? 2 -POL QPSK looks like the right solution 18

Customers want bigger pipes Why 100 G? 19

100 G - Things to know • 100 G is seen as the next big step for all vendors • First deployments are expected around 2010 timeframe • Given the lifetime of a transmission network, whatever is rolled out today should be 100 G ready • As the need for higher network capacities is there, a sitand-wait strategy is no option • Given the complexity of 100 G transmission, only vendors with solid 40 G knowledge & ASIC implementation have a realistic chance to get there in time 20

Advanced E/O Modem Introduction e. DC 100 • 100 G/ e. DC 40 10 G e. DCO • Electrical Tx based dispersion compensation Wraptor FEC • Better than +/- 50, 000 ps/nm • 9. 4 d. B of coding gain • Real-time Fully automatic • 3 d. B > RS-8 • Reach > 1000 Km • 2 -Pol QPSK 40 G • 10 Gbaud operation • Electrical CD and PMD compensation • +/- 50, 000 ps of CD compensation • 50 GHz OADM compatible • Electrical PMD mitigation • 50 GHz OADM compatible • Raman avoidance 2003 2005 2008 Future Embedding Transmission Complexity into Electronics 21 21

100 G Standards Update ITU Study Group 15 Q 6 Meeting – Oct 2007 • ITU determining next rate of OTN (OTU-4) to accommodate 100 Gb. E • OTU-4 rates considered: • 3 X 40 G -> 130 Gbit/s • 100 Gb. E -> 112 Gbit/s (most popular) • Decision on rate to be made end 2008 • Advanced modulation schemes considered to support the new rates: • Dual Polarization (Dual Pol) or Polarization Multiplexed QPSK, Duobinary, DQPSK, RZ-DQPSK. • Dual Polarization QPSK • Only format capable of 50 GHz spacing • Only format with 10 G-equivalent PMD tolerance • Only format that could transport both OTU-4 rate proposals 22 Expect Other Vendors to Move to Dual Polarization QPSK as Industry Moving in this Direction for 100 G

100 Gb/s Study Group Format Comparison NRZ-DQPSK DP-QPSK RZ-DQPSK Duobinary Bit Rate (Gbit/s) 112 130 Baud Rate [GBaud] 56 65 28 32. 5 56 65 112 130 Support of 100 GHz channel spacing Yes Prob Not Yes Yes Maybe No Support of 50 GHz channel spacing No No Yes (sim) No No CD Tolerance for 2 d. B OSNR penalty (ps/nm) +/-19 +/-14 18000 15000 +/-21 +/-15 +/- 23. 5 +/-17. 5 Max DGD tolerance for 1 d. B OSNR [ps] 6. 1 5. 3 27. 0 23. 0 7. 3 6. 3 2. 7 2. 3 OSNR tolerance for BER=1 e-4 19. 2 19. 8 16. 8 17. 4 18. 7 19. 3 21. 7 22. 3 2 d. B bandwidth of flat top filter for 2 d. B OSNR penalty +/-30. 2 +/-35. 2 +/-15. 4 +/-18 +/-28. 6 +/-33. 3 +/- 35. 9 +/-41. 7 Nortel Confidential 23 Only format capable of 50 GHz spacing Only format with 10 G-equivalent PMD tolerance Only format that could transport both OTU 4 rate proposals

OIF selects DP QPSK for 100 G 24

Agenda • Market trends • Main Services and Bandwidth drivers • Technology trends • Key network requirements • Ways to build a future-proof transmission network • DWDM transmission • 10 G - Still the baseline • 40 G - The next big thing • 100 G - On the horizon • Networking flexibility • • ROADMs - Photonic flexibility OTN - The successor of SDH (? ) L 2 - Embedded data capabilities Control plane - Gluing it all together • Network example Ro. Edu. Net • Outlook and summary 25

The future proof transport network Ingredients to Achieve All-Optical Agility ROADMs and OTN Seamless 10/40/100 G Electronic Dispersion Compensation 26

ROADM Applications and Drivers > Reconfigurable Optical ADM > Traditional networks require manual patching as OADM and increases automation and reduces pass-through requirements change OEO costs over time • Remote re-configurability > Line system optimization must be rebalanced with OADM reconfigurations • Optical branching • Router / DXC bypass ROADM > Automated System optimization & power balancing ROADM Rebalance and optimize as wavelength routing changes • All VOAs are electronic • All power control done remotely • No manual equalization Reconfigure with changing traffic requirements ROADM 27 • Automatic reconfiguration for nodal wavelength pass-through events – no manual patching required

ROADM Architectures 2 -Degree ROADM WSS Optical bypass traffic • Terminates wavelength services or passes them transparently through in the optical domain (no transponders / regenerators) • Connected to two fiber pairs (degree two) Multi-degree ROADM • Connected to at least three fiber pairs • Can lead to cross connections restrictions or scalability issues Add/drop and regen traffic Optical bypass traffic Add/drop and regen traffic Dir 2 Directionally Independent OADM • Guarantees non-blocking wavelength switching between fiber pairs • Allows any wavelength to be re-routed to any path on the network without manual intervention 28 Optical bypass traffic Dir 1 Directionally Independent Add/Drop Dir. N

Starplane http: //www. starplane. org/ 29 - 29

DAS-3 Network Overview University of Amsterdam (Uv. A) Media lab University of Amsterdam (Uv. A) VLE (Virtual Laboratory for E-science) Amsterdam Free University (VU) City ring? Cluster with blade PCs SURFnet CPL 8*10 G bandwidth Between each Node pair on CPL ring Leiden University (UL) Delft University (TUD) 30 - 30

Chosen Implementation - One band in SURFnet 6 Ring 1 (Green ring) allocated to DAS-3 - Dynamic switching using WSS and OME 31 But what about OTN…? - 31

The idea behind OTN > The G. 709 frame structure was defined to provide OAM for monitoring end-to-end services and protection capabilities for optical services, i. e. wavelength services > It supports the use of standard FEC and enhanced FEC when needed and inherently provides 3 R functionality > The frame structure was defined for 3 wavelength bit rate; 2. 5 Gbit/s, 10 Gbit/s and 40 Gbit/s (to match with the SDH clients) > Support of sub-wavelength services was not considered, as they could be provided by client layer networks SDH. IP ATM/FR MPLS SDH/PDH WDM 32 L 3: IP/MPLS L 2: Ethernet L 1: OTN

Reasons for the OTN Evolution > 10 Gb. E • Bit transparent transport of 10 GE (10 GBase-R) requires an over-clocked ODU 2. A number of proprietary implementations provide the required transparency. > Transparent transport 4 x 10 GE LAN over 40 Gbit/s • Requires a mapping into an over-clocked ODU 2 and multiplexing of them into a new over-clocked ODU 3. One further new function is needed. • The clock tolerance of ± 100 ppm requires a new multiplexing method of ODU 2 e • The use of the standard multiplexing method requires a new bit-asynchronous mapping of 10 GE > 40 GE could be mapped into the standard ODU 3 when transcoding is used. > 100 GE over a single wavelength requires a new ODU 4. > SDH supports transparent transport of 1 GE, but SDH will be switched off. Direct transport over the OTN requires a new sub-ODU 1/ODU 0. > The OTN must be timing transparent for Ethernet CBR signals in order to support Synchronous Ethernet 33

OTN Extensions Agreements OTUk/ Higher Order ODU NEW OTU 4 Lower Order ODU H-ODU 4 CBR Clients 100 Gbit/s 2 x 40 Gbit/s OTU 3 1 x ODU 3 OTU 1 1 x 10 GE 1 x STM-64 ODU 1 1 x STM-16 ODU 0 1 x 10 x ODU 2 1 x 4 x 2. 5 Gbit/s NEW 34 STM-256 ODU 2 x ODU 2 4 x 10 Gbit/s 1 x 40 GE 1 x 16 x OTU 2 1 x 1 GE classical OTU or ODU XXX new agreed OTU or ODU XXX 100 GE 1 x XXX new OTU, ODU not yet agreed standardized mapping or multiplexing new agreed mapping or multiplexing

Outlook on OTN Extensions OTUk/ Higher Order ODU Lower Order ODU CBR Clients L-ODU 4 OTU 4 y H-ODU 4 100 Gbit/s OTU 3 y 100 GE 1 x 2 x ODU 3 y 40 Gbit/s ODU 3 1 x 40 GE 1 x ODU 3 1 x ODU 2 y 10 Gbit/s OTU 1 35 10 GE 1 x STM-64 1 x STM-16 1 x 1 GE 10 x 16 x OTU 2 y OTU 2 1 x ODU 1 4 x STM-256 ODU 0 OTU 3 1 x ODU 2 1 x 2. 5 Gbit/s 4 x 8 x classical OTU or ODU XXX new agreed OTU or ODU XXX 1 x XXX new OTU, ODU not yet agreed standardized mapping or multiplexing new agreed mapping or multiplexing not yet agreed

The future proof transport network L 2 Awareness Ingredients to Achieve All-Optical Agility ROADMs and OTN Seamless 10/40/100 G Electronic Dispersion Compensation 36

MSPP Network Applications Multimedia Collaboration • Broadband Multiplexing • Ethernet Services Delivery MSPP Voice (Vo. IP) Storage/ILM Broadband Photonic Operations Interconnect 40 G • SAN Extension • Broadband Multiplexing • Ethernet Services Delivery • Infrastructure (ROADM vs OMX) • Ethernet Services Delivery Packet Optical Solutions are deployed in private builds, shared infrastructure and managed services solutions. 37 37

Network Applications - SAN Extension Transactions performed locally Metro Network OM 5000 Database / Storage Array Data stored / backed up Remotely Transaction OME 6500 Database / Storage Array Transaction Addresses SAN Extension with requirements of intermediate multiplexing of services 38 38

Network Applications - Ethernet Services OME r Fibe Copper/ fiber Fiber Metro/WAN Copper OME 61 x 0 MPLS Core DWDM/ SONET/ Ethernet OC 3/12/48 Nx 2. 5 G OME 6500 OM 3500 Ethernet VPN solutions on any of the converged layers 39 39

The future proof transport network Control Plane L 2 Awareness Ingredients to Achieve All-Optical Agility ROADMs and OTN Seamless 10/40/100 G Electronic Dispersion Compensation 40

Optical Network Automation Objectives Optical Control Plane Optical Layer > Network Topology Discovery and Awareness > Automated Service Activation • Can be “Client” or “Operator” driven > OEO & OOO technology provides economical flexibility • OEO for Service adaptation, network adaptation and monitoring • OOO for photonic flexibility / re-configurability > Leads to Network protection / restoration > Potential for IP / Optical inter-working via GMPLS signaling 41

Considerations when Control is Enabled Optical Control Plane Optical Layer > Control Plane in an Optical Network enables: • • • Automated service activation in optical layer Network awareness resource status and utilization Rapid identification / correlation of fault / resource / service Optical protection and restoration Ability to add a new wavelength automatically without impacting existing network Understanding the viability of the end-to-end wavelength path is critical 42

Mesh Restoration S 1 I 1 Failure notification I 2 P 1 D 1 P 2 • Automatic Restoration recovers traffic following a path failure • For traffic not protected by the Transport Plane (e. g. 1+1) • For backup restoration (e. g. 1+1 secondary) • Dynamic restoration scheme for best survivability and efficiency • Control plane learns location of failure in the signaling notification, computes next best route based on feedback information and re-routes each connection • No pre-computed/pre-assigned restoration path/bandwidth for higher bandwidth efficiency. Mesh Restoration will recover from multiple failures as long as b/w is available for restoration • Restoration performance is fundamentally unpredictable and non-deterministic, therefore restoration times are typically slower i. e. in the range of secs • Example: For 1+1 Path Protection Co. S, Automatic Restoration may be optionally used to restore 1+1 path protection after initial failure • When a working connection fails, traffic is protection switched to protecting connection within 50 ms by Transport Plane. • CP then re-creates (restores) the failed working connection to return the Co. S back to the 1+1 Path Protected state. 43

The future proof transport network Ingredients to Achieve All-Optical Agility Control Plane L 2 Awareness ROADMs and OTN Seamless 10/40/100 G Electronic Dispersion Compensation 44

Agenda • Market trends • Main Services and Bandwidth drivers • Technology trends • Key network requirements • Ways to build a future-proof transmission network • DWDM transmission • 10 G - Still the baseline • 40 G - The next big thing • 100 G - On the horizon • Networking flexibility • • ROADMs - Photonic flexibility OTN - The successor of SDH (? ) L 2 - Embedded data capabilities Control plane - Gluing it all together • Network example Ro. Edu. Net • Outlook and summary 45

Ro. Edu. Net Next Generation Network • 4. 238, 8 km fiber • 57 locations • 22 OME 6500 • 18 ROADM sites “…to offer the participants - universities, high schools, cultural, scientific and research nonprofit institutions - the means to communicate with each other…” 46

CAREI SATU MARE BAIA MARE ILVA MICA ILVA NUCA VATRA DORNEI PASCANI SUCEAVA IASI NOC MARGHITA DEI JIBOU CIUCA VASLUI MURES NOC RAZBOIENI TGTG MURES NOC CLUI NAPOCA BACAU ORADEA TECUCI NOC ALBA IULIA RUPEA TEIUS GALATI TEIUS CHISINEU CRIS ALBA IULIA SAVARSIN ARAD TEIUS NOC COPSA MICA FOCSANI DEVA SIBIU RM. VALCEA BRASOV DEVA CAINENI BRAILA BRASOV PITESTI NOC TG JIU PLOIESTI FAUREI TIMISOARA TARGOVISTE PETROSANI BUZAU NOC ROADM TRAFFIC SITE NETWORK DIAGRAM WSS module INTERMEDIATE SITE BUCURESTI CRAIOVA ROSIORI CIULNITA BUC 47 NAT FETESTI CONSTANTA

OME 6500 Network Convergence Versatile L 0/L 1/L 2 Convergence Platform L 2 Termination ROADM ü RPR ü L 2 SS for packet aggregation ü Termination of DS 1/E 1, DS 3/E 3 on L 2 SS for Off-net ü network agility ü Single add and drop granularity ü Restoration 40/100 G Adaptive Optical Engine ü Innovative technology for simpler network deployments No hard hats ü Smooth migration required 10 40 100 G ü>> 1000 km reach without REGENs MSPP Transponders ü 2. 5 G to FC 1200 ü Multiple protection options ü OTN-based transponders Customer Network ü SONET, SDH, J-SDH ü International Gateway ü Next. Gen DCS ü LO and HO crossconnects ü Full range of transport services VT X-Connect OC-n Port Card 5 G TMUX DS 3 Term OTSC 1 + 1 L i n e STS Mapping DS 1 VT Term Mapping & BP driver VTU Optics 24 x DS 3/EC-1 Port Card 1 2 BP driver VTU … global platform with one software load… any card, any service, any chassis … 48 DS 3/ EC-1 Term 24

Optical Multiservice Edge family OME 6500 Double Decker OME 6150 OME 6500 ANDA OME 6110 OM 5065 OME 6130 OME 1110 Demarc 49 SONET / SDH CPE SONET / SDH OME 6500 Family

Possible Network Migration to 40/100 Gbps ROADM Terminal OME 6500 ROADM 50 GHz System Terminal OME 6500 To Add 40 Gbps Wavelength: 1. Insert e. DC 40 line and 40 G client cards in each OME 6500 terminal shelf 2. Connect fiber from e. DC 40 card into existing long haul or metro line system 50 3. Connect client signal to 40 G Client Card

Agenda • Market trends • Main Services and Bandwidth drivers • Technology trends • Key network requirements • Ways to build a future-proof transmission network • DWDM transmission • 10 G - Still the baseline • 40 G - The next big thing • 100 G - On the horizon • Networking flexibility • • ROADMs - Photonic flexibility OTN - The successor of SDH (? ) L 2 - Embedded data capabilities Control plane - Gluing it all together • Network example Ro. Edu. Net • Outlook and summary 51

Target Packet/Optical Network architecture MPLS Services Ethernet Services (RFC 2547 VPN, PWs etc. ) (E-LINE, E-TREE, E-LAN) PBB / PBT / PLSB L 3 VPN MEF UNI GMPLS for L 2 ITU-T interlayer DWDM / OTN OIF UNI GMPLS for L 0/1 ITU-T interlayer PCE ASON/GMPLS Architecture TDM Services (SDH, Sonet, PDH) Ethernet Services (E-LINE, E-TREE, E-LAN) 52

Conclusion • Bandwidth demand keeps on growing • Network flexibility is key • Optical transmission networks have to be ready for future upgrades to higher bitrates • ROADM, OTN, embedded L 2 -features and control planes will lead to a new level of flexibility Ro. Edu. Net’s Next Generation Optical Transport Network is the perfect base for current and future services 53