Seminar to Caltech professors and graduate students about innovations in AT&T Labs; March, 1999

1. ACCESS TECHNOLOGY
innovations in AT&T Labs
Xiaolin Lu
March, 1999 @ Caltech
AT&T Labs

2. AT&T Labs - RESEARCH
Larry Rabiner
Vice President
Laboratories
Information Communications Information Speech & Image Network Services Networking and
Systems & Services Infrastructure Sciences Processing Research Distributed Systems
Research Research Research Services Research
Chief
Scientist
Ron Brachman Ted Darcie Rob Calderbank David Berkley Dave Belanger Hamid Ahmadi
Specification & Algorithm
Artificial Intelligence Wireless Systems Mathematics and Speech Processing Innovative Services Research
Principles Nelson Sollenberger Cryptography Software & Technology David Unger Michael Merritt
Michael Kearns Andrew Odlyzko
Broadband Wireless Bishnu Atal Information Systems Network Mathematics
Machine Learning and Systems Communications & Analysis Research Sandy Fraser
Information Retrieval Paul Henry N. Seshadri Rich Cox Ken Church Albert Greenberg
Fernando Pereira
Wireless Statistics Candace Kamm Software Systems Systems & Networking
Human / Computer Communications Daryl Pregibon Phong Vo Resources Chief
Interface Larry Greenstein Jay Wilpon Frank Pirz Scientist Emeritus
Julia Hirschberg Algorithms and Large-Scale
Broadband Access Optimization Image Processing Programming Networking Research
Secure Systems Adel Saleh David Johnson Software & Technology Bjarne Stroustrup Charles Kalmanek
David Maher
Lightwave Networks Algorithms and Barry Haskell Information Services Distributed Systems
Online Platforms Bob Tkach Distributed Data John Denker Michael Merritt (Acting)
Gregg Vesonder Joan Feigenbaum Yann LeCun
Communications Customer Information Database Research
Adaptive Information Technology Behzad Shahraray Mike Wish H. V. (Jag) Jagadish
Services Robert R. Miller
Larry Jackel Information Consumer Electronics Ron Graham
Visualization Ed Chen
Stephen North
Broadband Services
Norm Schryer
Unofficial 10/1/98

3. Lightwave Technology in AT&T Labs
Long-Haul Access
WDM/xDFA Optical
ORAN PON HFC
High Capacity Networking

4. ACCESS ?
Toll

5. INTERDEPENDENCY
VOICE VIDEO
 Communication  Entertainment
Services  Point-to-point  Broadcast
 “On-Demand”  Always there
Billing  Usage-based  Flat rate
 Narrowband/  RF Broadband
Transport
baseband
 Point-to-point  Point-to-multipoint
Network  Coax
 Twist pair

6. ACCESS ENVIRONMENT
LEC Cable
Voice Video
 Narrowband  Broadband
 Circuit Switch  Broadcast
 Dedicated  Shared
 Synchronous  Always there
 Usage-based  Flat-rate
billing billing
DBS/LMDS
Cellular /MMDS

7. A HISTORICAL PERSPECTIVE
LEC-
LEC-Centric
- Twist Pair
- Circuit switch
- Voice & low speed data
 Regulated
- Monopoly
 DLC
- Access charge
 ISDN  Voice-oriented
Voice-
Point-to-
Point-to-point
Fiber-To-The-
Fiber-To-The-Home

8. CHANGING PERSPECTIVES
Telco-Centric Mosaic
RF and DSP
 Wired/wireless
 Modem technology
 Wired  Compression/codec  Broadcast
 Switched /narrowcast
 Broadband SCM Lightwave  Broadband
/narrowband
 Video + voice + data
 Reliable + low-cost
 Synchronous Data Applications  Asynchronous
 Internet
 Piggy-back
 Monopoly Legislation  Competition
 Time-to-market
Time-to-
 Telecom Reform  Techno-economic
Techno-

9. Ten Year Performance Improvement
WAN Bandwidth 2 2000
Processor Power 2 1000
Router Engine Performance/Price 1 1000
Internet Traffic 2 1000
Bandwidth to Homes 3 13
Sources: 1. Business Communications Reviews, Sept. 1997
2. AT&T
3. Dataquest

10. NETWORK TRAFFIC
1000
Voice
800
Data
Traffic (Gb/s)
600 Total
400
200
0
1990 1995 2000 2005 2010
K.G. Coffman and A. Odlyzko
AT&T Labs

11. ACCESS REVOLUTION
 Interconnect routers with high-speed Regional / Metro
(OC3  OC48  …) ports Networking
 Migration of corporate LAN bandwidth Broadband
to consumers (1  10  100 Mb/s) Wireline Access
 Intense competition between broadband Broadband Fixed
Wireless Access
access providers
 Strong demand for mobile/tetherless 3rd-Generation
Mobile Wireless
voice and emerging data capabilities

12. END-TO-
END-TO-END NETWORKING
National
Metropolitan/Regional
Local
Wireless Cable Copper Pair Fiber Other Access

13. Optical Regional Access Hierarchical Network
Backbone Network Feeder Ring
Multi-fiber WDM Ring
POP Configurable Nodes
Directly Connected
Customer Distribution Network
Egress Passive, WDM
Node Rings, Trees, Buses
Access Single-Wavelength Ring
Node Electronic TDM
Feeder
Ring
Passive Access Access
Distribution Node Node
ADM
Tree Single- ADM
Directly Wavelength
ADM Connected Ring
Single- ADM Passive ADM
Wavelength Distribution
ADM Ring
Ring
ADM
- Customer Premises
- Passive Splitter or Add/Drop
ADM - SONET or ATM Add/Drop Multiplexer

14. COMPETING WIRED ACCESS SYSTEMS
FIBER-TO-THE-
FIBER-TO-THE-CURB (FTTC)
ONU  Switched bandwidth
CO HDT  Cost? Analog video?
HYBRID FIBER COAX (HFC)
 Segmented broadcast
bus; FDM/TDM
CO FN  Low-
Low-cost broadband
 Upstream?
PASSIVE OPTICAL NETWORK (PON)
ONU  Passive outside plant
CO HDT  WDM upgrade
 Cost? Analog video?
Splitter/WDM

15. ACCESS OPTIONS
STRATEGY MENTALITY PREFERRED SYSTEMS
Operation
Saving HYBRID FIBER/COAX
REBUILD
NEW Secured
FIBER TO THE CURB
BUILD Switching
Embedded Future
PASSIVE OPTICAL NETWORKS
Proof
New Entry
Low-
Low-cost
Entry FIXED WIRELESS
EARLY
MARKET
ENTRY Utilize
Existing  CABLE MODEM
Network  xDSL

16. EVOLVING
Transport
- Lightwave
- RF
Network
 Upgrade - Distribution
- Quality - Trunk
 Broadband
SDV  AM-VSB
- Reliability
- Bandwidth Terminal
- Upstream
- Modem
- Protocol
Operation
- Monitoring
- Management

17. EVOLUTION AND REVOLUTION
Need
Architecture
More Evolution/
Revolution
Digital Cable
Services
Modem
Quality
Reliability
HFC
Linear Lightwave RF & DSP Low-cost lightwave Technology
and WDM

18. Coax to HFC
HE

19. Coax to HFC
FN
FN
HE
FN
 Increase transport capability
 Improve quality and reliability
 Challenge for linear lightwave

20. CHALLENGES OF AM-VSB TRANSMISSION
AM-
XTR RCV
Requirements:
CNR 52dB 47dB
CTB/CSO -65dBc -55dBc
Saleh
Limit
Challenges:
CIR
- Linearity - Dispersion - Linearity
- Noise - Reflection - Noise
- Chirp - Others
OMD

21. RF MODEM TECHNOLOGY
I I
10110 10110
DMUX Carrier Carrier MUX
S S Recovery
900 900
Q Q
Modulation Spectral Efficiency CNR
Technique Theory Practical (10-8)
QPSK 2 1.2 - 2 15
16-QAM 4 2.5 – 3.5 22.5
64-QAM 6 4.5 – 5 28.5
AM-VSB 47

22. Performance of An Uncooled FP Laser
 No TE cooler, no isolator  60 QPSK channels at 2Mbps/ch
 ~ $100  Temperature: 200C ~ 800C
-2
Theory
Room Temperature
-3 T>80C
Reflection
log(BER)
-4
-5
-6
-7
-8
-9
-10
2 4 6 8 2 4 6 8 2
0.1 1 10
OMD(%) S. Woodward, G. Bodeep, OFC’95

23. New Transport Opportunities
DSP Multimedia
 MPEG-2,
MPEG- etc Digitization
RF Modem Digital SCM System
 BW efficiency
 Robustness  Broadcast/narrowcast
digital TV
Digital RF
Transmission S  Interactive video
RF/(wireless)  2-way communication
 Front-end
Front-  Voice, data, etc
 Wireless backhauling
Linear Lightwave Low-
Low-cost
 Maturity Lightwave

24. EVOLUTION AND REVOLUTION
Need
Architecture
More Evolution/
Revolution
Digital Cable
Services
Modem
Quality
Reliability
HFC
Linear Lightwave RF & DSP Low-cost lightwave Technology
and WDM

25. CHALLENGES
HE FN
HE FN
HE
FN
Analog Emerging
TV Services
5 50 500 750 1G
 Bandwidth Capacity: 5-40MHz/1000s HHP upstream
 Transport Integrity: Ingress noise, dynamic range
 103-to-1 Architecture:
to- Centrally-
Centrally-mediated MAC

26. SOLUTIONS
Bandwidth UPGRADE
Capacity  Fiber Node
Network
Segmentation
 DWDM Trunk
Transport
Integrity DOCSIS
 High level
modulation Modem
 Centrally-
103-to-1 mediated MAC
Architecture

27. Fiber Node Segmentation
HE FN
1,200 Homes
 Long cascade coax bus shared by many users (1000s)

28. Fiber Node Segmentation
300 Homes 300 Homes
HE FN
300 Homes 300 Homes
 1,200 HHP/FN with 300 HHP/Bus

29. DISTRIBUTED HEAD-END
HEAD-
HE
FN
Primary Primary
Hub
HE
Ring
FN
HE
 Operation complexity
 Cost of CMTS at lower take rate

30. DWDM TRUNK
SH
FN
Primary Primary
Hub
SH
Ring
FN
SH
 DWDM transport for end-to-end transparency
end-to-
 Route diversity for service protection
 Consolidate high-end terminals (CMTS)
high-

31. DWDM TRUNK
Primary Hub Secondary Hub
1.3mm
XTR
Coarse
WDM
l
1 x 8 DWDM
1 x 8 DWDM
l
l
. .
.
.
. . Fiber Node
1.5mm RCV
RCV l
1 x 8 DWDM
1 x 8 DWDM
RCV l
RCV
.
. .
.
. .

32. MODERN HFC NETWORK
SH FN
Primary Primary
Hub
SH
Ring
SH FN
DWDM Transport Segmentation
End-to-end Transparency 4X capacity

33. What If We Succeed?
 Bandwidth exhaustion  Transport integrity
 Take rate and multiple lines
 New services
 User behavior
 Performance
10000
1000
Delay (ms)
100
Life cycle cost
10
v.s.
1
10 20 30 40 50 60 70 80 90 100 Front-
Front-end cost
Users

34. ARCHITECTURES
Tree-and-Branch
 Broadcast FN
 Cascaded
???
Cell-Based
 Narrowcast RN
 Clustered

35. Mini Fiber Node (mFN)
mFN
FN mFN
Primary Primary
Ring Hub
mFN
FN mFN
DWDM
Analog Digital
TV TV
5 50 500 750 1G
 mFN overlay for a cell-based digital platform
cell-
 Multi-
Multi-purpose infrastructure
 50-
50-100 times more clean two-way bandwidth
two-
 Distributed MAC protocol with better performance

36. Early Version of an mFN Prototype

37. Line extender mFN

38. DELAY COMPARISON
1000
100
Average delay (ms)
10 mFN-NAD
CM
1
0.1
mFN-NAD Cable modem
0.01
10 20 30 40 50 60 70 80 90 100
Number of active users

39. Mini Fiber Node (mFN)
mFN
FN mFN
Primary Primary
Ring Hub
mFN
FN mFN
DWDM
PCS
Analog Digital
TV TV
5 50 500 750 1G 2G
 Multi-
Multi-service platform

40. EVOLUTION
Demand
Bandwidth per Customer
Take Rate
Applications
User Behavior
Push Fiber
Deeper
Split Nodes
Higher RF
Efficiency
Time

41. Fiber Optics for Cable
108 FTTH
107
FTTC (102 HP/node)
106
Fiber termination
105 FSA (103 HP/node)
100 ch AM
104
103
102
40 ch AM
86 92 00 06
Broadcast Two-way Broadband

42. PASSIVE OPTICAL NETWORKS
CONVENTIONAL TDM PON
3 2 1 1
LASER 3 2 1 PS •
• • BROADCAST BASEBAND
PASSIVE • COMMERCIAL TECHNOLOGY
•
SPLITTER
SUBCARRIER (SCM) PON
1
•
LASER PS •
• BROADCAST PASSBAND
•
• AVAILABLE TECHNOLOGY
1 2 3 1 2 3
RF RF
WAVELENGTH MULTIPLEXED (WDM) PON
1
1
LASER WDM •
• • VIRTUAL POINT TO POINT
1
2 • EMERGING TECHNOLOGY
3 •

44. WIRELESS ACCESS SYSTEMS
Cellular
- Narrowband/Mobility
- High Cost/Poor quality
Compete to
wireline (voice) Compete Enhanced
Services
to cable
Wireless Loop Broadcast PCS
Broadband
Internet/
WWW
Universal
Coverage
Data
- Narrowband: pagers LEO Satellite
- Broadband

45. Broadband Wireless Field Experiment
Modem
Internet Bell Atlantic
Intranet
Modem
Head-End
Platform
Servers
NSL ‘Switching Center’ Navesink Tower
Modem
PURPOSE: PSH Office
•Explore and extend technology
•Demonstrate proof-of-concept

48. Wireless Propagation Data

49. Wireless: Tricks to Increase Capacity & Peak Rate
What How
Advanced Coding Turbo, layered or space-time codes
Adaptive Coding Adjust code rate to maximize performance
Variety of techniques to combat fading and
Diversity and Equalization
multipath dispersion
Smart Antennas Adaptive processing using antenna arrays
Dynamic Resource Allocation Spectrum or timeslots assigned only as needed
Minimize transmit power through measurement
Power Control
and feedback
- Orthogonal FDM
Advanced Modulation - Wideband-CDMA
- M-PSK
 Introducesignificant complexity in required digital signal processing
 Rapid progress driven by increases in DSP horsepower (1000X in 10 years)

50. SUMMARY
 Emerging technologies, new service opportunities and
competition are driving wide variety of access
infrastructure choices
 Cost of photonic components and embedded base keep
electronic options (copper pair, coax) for last mile (HFC,
FTTC)
 Desire for passive future-proof outside plant favors PON
 Fast time-to-market and new service opportunities make
wireless alternatives more attractive

51. “There is no wrong technology,
there are only wrong
assumptions.”