A » switched photonic network for the new transport backbone of Telecom Italia

GRUPPO TELECOM ITALIA

Photonics in Switching 2009Tirrenia, September 17 2009

A λ switched photonic network for the new transport backbone of Telecom Italia

Piergiorgio Pagnan, Marco Schiano

Transport & OPB Engineering

Transport & OPB Innovation

Telecom Italia Transport & OPB InnovationMarco Schiano

Summary

► Evolution drivers and requirements

► Network structure and dimensioning

► Network scalability

► Toward 100 G networking

► Future evolution and research challenges

Photonics in Switching 2009A λ switched photonic network for the new transport backbone of Telecom Italia

Marco Schiano, Transport & OPB Innovation 3

Why a new Photonic Backbone?

► To cope with a remarkable traffic increase

► from the domestic networks (especially the IP backbone)

► from International networks

► from the emerging λ wholesale market

► To decrease costs (both CAPEX and OPEX)

► To enhance reliability for critical services

► To reorganize the transport backbone into a single easily manageable platform, switching over multiple legacy DWDM systems



The domestic client networksIP backbone architecture

OPB: Optical Packet Backbone

M I

PD

TSBS

BO

TO

GE

FI

PA

RM

NA

BA

SV

AL

BGCO

VRVE

BZ

MORI

PI

ANPG

PE

CA

TA

CZ

CT

NL

M I

PD

TSBS

BO

TO

GE

FI

PA

RM

NA

BA

SV

AL

BGCO

VRVE

BZ

MORI

PI

ANPG

PE

CA

TA

CZ

CT

NL

TO

AL

VRVE

RM 2RM 1

MI 1MI 2

RM 2

CT

PA

RM 1

NA

RM 1RM 2

VRVE

MI 1MI 2

BO

PC

RM 1RM 2

SS

CA

SS BA

TANA

BO

PI FI

PC

RM 1RM 2

PE

AN

PG

RM 1

XX

► CRS 1 Tera-routers in the core

► 10 Gbit/s POS interfaces for all links

► 40 Gbit/s POS interfaces in the core

► ASON meshed network

► SDH cross-connects and DWDM links

► Control Plane, centralized routing

► 2.5 Gbit/s SDH ring architecture

► Today used for structured VC4 services

► Excellent reliability (MS-SPRing)



Carrying international traffic through Italy

MedNautilus

Telecom Italia SparklePan-EuropeanBackbone

► Traffic originated in far and middle east is conveyed to Sicily by submarine systems

► It has to be delivered to northern Italy where the Telecom Italia Sparkle Pan European Backbone PoPs are located



Opportunities of new photonic technologies

TransparentOptical Tunnel CP CP

CP

CP

CP

CP CPCP

Ultra Long-HaulDWDM

Multi-degreeROADM

► Fewer regenerators

► CAPEX savings

EnhancedGMPLS

Control Plane

► End-to end provisioning

► OCh protection and restoration

► OPEX savingsProtected/Restored

Path



The Photonic Backbone structure

► Network diameter: 2400-3100 km (working-protection paths)

► Maximum number of hops: 11

► Nodal degree: 2÷5 (av. 3.1)

► Technology:

► ∼40 λ switching nodes based on ROADMs

► ∼60 ULH DWDM systems with 80 lambdas

► G.655 and G.652 fibers

► 10 and 40 Gbit/s optical channels (OCh)

► Ready for 100 G transmission

∼1000 km

∼500 km

Tentative scheme

of the new Backbone



Preliminary network dimensioning

► Traffic load rough forecast in 2011:

► 542 OCh (mixed 10 and 40 Gbit/s)

► ∼ 100 OCh are used for traffic protection

► Maximum number of hops: 9

► Maximum OCh length: ∼ 1600 km

► 6 Tbit/s total capacity

► No serious issues of lambda congestion at least with 2011 traffic estimates

► A double fully disjoint protection path is not available for all OCh due to cable topology limitations



Network scalability raw estimation

► OCh network capacity: ∼ 1300 OCh(estimation based on present traffic pattern)

► Bandwidth capacity ranges from 13 to 130 Tbit/s depending on OCh bit rates

10 G

Tbit/

s

13

52

130

40 G

100 G

Total network capacity: 1300 OCh

►2011 network utilization (542 OCh, 6 Tbit/s):

►∼ 30% OCh exploitation

►∼ 11% bandwidth exploitation (referred to fully 40G network)



Network scalability analysis

► Increasing OCh number may be limited by fully loaded DWDM links

► Adding new DWDM systems “in parallel” is the easiest way to upgrade the network, provided that new ROADM line ports and new fibers are available

New DWDMsystem

CP CP

CP

CP

CP

CP CPCP

CongestedPathUnique cable

New line porton ROADMs



DWDM systems as a function of increasing OCh

Number of Systems ( 80 Lambda @ 40G)

0

10

20

30

40

50

60

70

80

90

Nominal Demand(542 λ)

+20 % (651 λ) +50 % (813 λ) +100 % (1084 λ) +200 % (1626 λ)



Nodal degree as a function of increasing OCh

0

1

2

3

4

5

6

7

8

9

BAR

CAT

CTZ

LAM

MAZ

NAP

NO

L

PAR

PES

RM

I

RM

S

TAR

ALE

ANC

BOP

BLZ FIR

GEN MIB

MIM

MO

D

PDS

PRM

PGA

PES

PCA

PIS

RIM

RM

I

RM

S

VRS

SAV

TOR

TRT

TRE

TRI

UD

I

VEN

Nom+20 %+50 %+100 %+200 %



CAPEX evaluation for OCh increasing: common parts

Icremental cost for common parts (ROADM and OLA)

0%

5%

10%

15%

20%

25%

30%

35%

40%


+20 % (651 λ) +50 % (813 λ) +100 % (1084 λ) +200 % (1626 λ)



CAPEX evaluation for OCh increasing: whole network

Incremental CapEx for the whole network

0%

20%

40%

60%

80%

100%

120%

140%

160%

180%


+20 % (651 λ) +50 % (813 λ) +100 % (1084 λ) +200 % (1626 λ)

REGENERATORTRANSPONDEROLAROADM



Network lifetime expectation

► Initial network deployment foreseen in 1Q 2010

► Long term traffic estimates are unreliable, however:

► If the annual traffic growth rate is 25%, the total network load will be around 55 Tbit/s in 2020

► Even higher traffic growth rate, up to 40% per year, can be sustained by the new photonic backbone in the next decade



When 100 G will become economically viable?

Assumptions on transponders cost ratio: 40G/10G=2.5, 100G/40G=2

Relative Network Cost (reference: traffic of 6 Tbit/s and Lambda @ 10 Gbit/s)

0

1

2

3

4

5

6

7

8

9

λ @ 10 Gbit/s λ @ 40 Gbit/s λ @ 100 Gbit/s

60 Tbit/s

30 Tbit/s

18 Tbit/s

12 Tbit/s

6 Tbit/s



When 100 G will become mandatory?

Maximum nodal degree required

0

5

10

15

20

25

30

35

6 Tbit/s 12 Tbit/s 18 Tbit/s 30 Tbit/s 60 Tbit/s

λ @ 10 Gbit/s

λ @ 40 Gbit/s

λ @ 100 Gbit/s



Energy savings and other operational benefits

► Compared to transport on point-to-point DWDM systems, energy savings range between 20 and 30%

► Energy saving is mainly due to the regenerator number reduction, while ROADMs power consumption is very low

► Other important benefits are:

► Remarkable spare parts reduction (due to fewer regenerators);

► ~40% circuit creation cost reduction;

► Opportunity of relocating the circuits of legacy networks on the new backbone simplifying the transport in the backbone



Expected improvements and research open issues

► Node architecture for scalable colorless-directionless add/drop

► Coexistence of 10 G intensity-modulated and 40 G phase-modulated OCh

► 100 G transmission reach

► Control Plane

► Restoration speed

► ….



Node architecture

► Complexity of Colorless Directionless node architectures

P. Roorda, B. Collings “Evolution to Colorless and Directionless ROADM Architectures”, Paper NWE2, OFC 2008

► Scalable node architectures for traffic add-drop have already been proposed

► New optical components are required



10 G and 40 G OCh coexistence

► In a first deployment stage intensity modulated 10 G and phase modulated 40 G OCh will coexist in the network

► Appropriate launch power limitation must be set for 10 G OCh in order to avoid serious impairments on 40 G OCh

► This may limit seriously the 10 G reach

► Smart solutions such as 10 G and 40 G allocation in separate bands may be studied



100 G transmission reach

► Basic requirement for 100 G system: at least 1000 km reach, hopefully 1500

► Alfiad, M.S. et al., 111-Gb/s Transmission Over 1040-km Field-Deployed Fiber With 10G/40G Neighbors, Photonics Technology Letters, IEEE, Volume 21, Issue 10, May 15, 2009, Pages: 615 - 617



Control Plane and restoration speed

► Control Plane of Transparent Photonic Networks are designed as evolution of the of opaque networks CPs’(especially SDH) adding RWA functions

► The maturity of this technology has not been proven in the field yet

► Restoration time is still seriously limited by:

► Amplifier power and gain adjustment mechanisms;

► OCh power equalization;

► Optical or electrical dispersion compensation devices



Conclusions

► The new photonic backbone provides flexibility and bandwidth that seem sufficient at least for the next 10 years

► It will represent the fundamental optical transport layer supporting all legacy and possibly new client networks

► Future network enhancements will impact on the colorless-directionless node architecture, on the OCh bit rate and on the restoration speed



Acknowledgments

We gratefully acknowledge the fundamental contribution of several colleagues:

Sergio AugustoMaurizio BartoliValentina BriziCarlo CavazzoniAndrea Di GiglioGiuseppe FerrarisMarco QuagliottiAlberto Rossaro

A » switched photonic network for the new transport backbone of Telecom Italia

Documents

Transcript of A » switched photonic network for the new transport backbone of Telecom Italia