Adding Fiber to Your Access Network Diet

1
Adding Fiber to Your Access Network Diet
Ed Templeman Marketing Director
2
Conventional Wisdom
You Can Never Be Too Rich, Too Thin, or Have Too
Much Bandwidth
3
Perspective

• Why More Bandwidth Is Better
• YouTube consumes more bandwidth todaythan the
  entire Internet did in 2000
• It took 200 years to fill the Library of Congress
• 57 million documents
• 29 million books
• 12 million photos
• Worldwide, an equal amount of digital info is
  generated almost 100 times per day
• Deloitte Consulting predicts thatpeer-to-peer
  video demands may saturatethe Internets main
  backbones this year

4
Prepare for the Great Flood
The Exaflood is Coming
The Exaflood is Here

• One Exabyte 1.074 BILLION Gigabytes
• 2 Exabytes Total volume of information generate
  d worldwide in 1999
• 2 Exabytes Total volume of data currently
  handled worldwide by the Internet

every two hours
5
A Giant Sucking Sound

• The Biggest Bandwidth Absorbers
• IPTV standard and especially HD
• Internet content-on-demand (The wave of the
  future)
• TV Shows Movies TiVo/Amazon, Apple, Walmart
• ABC, NBC, CBS, FOX, etc.
• YouTube, other user generated content sites
• Peer-to-peer Internet video-on-demand (Also the
  future)
• Joost, Vuze, VeohTV, Babelgum, etc.
• Interactive Community-building
• Online gaming
• Xbox, Playstation
• Massively multi-player games

6
Money to Be Made

• The Biggest Absorbers Biggest Opportunities
• IPTV - 5.79 billion by 2011
• 60 of communications execs see significant
  revenue from IPTV within 3 years
• 28 confident of significant revenue from IPTV
  within 12 months
• Peer-to-peer other Internet video - 4.4
  billion by 2010 (Ad supported)
• Video-on-demand - 4.7 billionby 2012
• Online gaming OENs Fision
• Houston Today
• Hard-core gamers paying 70/mo for10 Mbps or
  159/mo for 20 Mbps FTTH
• Faster speeds more battles won

7
SD or HD, We All Scream for TV

• HDTV Adoption is Accelerating
• Today - Nearly 15 of American homes have HDTV
• 2010 Nearly 66 or 78 million homes have HDTV
• Many will have 2-3 HDTV sets
• HD SDTV from Internet IPTV
• Projected market requirements
• Average demand will require 2 HD 2 SD streams
• Also voice data
• Needs 20 Mbps today 40 Mbps by2020 Beyond?
• IPTV over xDSL falls short
• Not enough to deliver more than one HD
  stream/home
• MPEG-4 wont help

Sources Yankee Group, 2005 LightReading
Insider, 2007
8
Changing Bandwidth Envelope
Peak Bandwidth Needs (Mbps)
Source Broadband Properties, Sept 2006
9
SeeingThe Light
10
Seeing the Light

• Forces Driving Fiber Adoption
• Upgrade cycle
• Only 24 of surveyed IOCs CLECs hadupgraded to
  next-gen networks
• Another 50 plan upgrades in next 1-2 years
• Demand for IPTV, VoD, etc. is motivation
• Economics
• Greenfields refurbs
• Maintenance advantages
• Regulatory relief from UNE rules
• Copper costs giving fiber the advantage
• Forces Hindering Fiber Adoption
• Up-front costs for equipment and construction
• Plant and electronics with residual book value

Source RW Baird Survey
11
Seeing the Light

• Its All About Making the Right Choices
• Technologies
• Passive Optical Network (PON) GPON, BPON
• Active Ethernet
• Advantages and issues with each
• Market Darwinism economics is determining
  winner
• Video service delivery methods
• In-band IPTV
• RF overlay
• Broadband (Real Internet) TV
• Network architectures
• Homerun (P2P) / Centralized split
• Partially distributed split
• Fully distributed split

12
Technology Contenders

• Passive Optical Network
• Characteristics
• Point-to-multipoint Minimized, shared fiber
  optical electronics requirements
• Active electronics only at CO and customer
  premises
• Only fiber and passive splitters in OSP
• Minimizes locations needing AC power cooling
• Decreasing per subscriber expense
• Future scalability
• Types of PON
• BPON (Broadband PON)
• 622 Mbps down 155 Mbps up 132 Split
• Most widely deployed, but older technology
• GPON (2.4 Gbps older 1.2 Gbps)
• 2.488 Gbps down 1.2 Gbps up 132 or 164 split

13
Generalized PON Architecture

• Benefits
• Low-cost for high total bandwidth matches video
  broadcast traffic patterns
• Flexibility in outside plant topology
• Challenges
• More complex outside plant topology
• Choices GPON, BPON, APON, EPON

Connectors (ONT)
Drops
Passive Optical Splitter
Central Switch
Feeder
OLT
14
Technology Contenders

• Active Ethernet
• Characteristics
• Point-to-point Dedicated 100 Mbps 1 Gbps per
  subscriber
• Active electronics at CO, remote cabinets,
  customer premises
• Lower cost optics Greatest number needed
• Dedicated optical electronics home run fiber
  for each subscriber
• May aggregate in active remote due to distance
• Constant per subscriber expense
• Future scalability

15
Generalized Active Architecture

• Benefits
• Dedicated bandwidth per subscriber
• Simple, point-to-point topology
• Challenges
• Cost Each subscriber requires a separate pair of
  optical transmitters/receivers
• Limited deployment options

Connectors (NID)
Drops
Central Switch
16
Battle of The Bandwidths

• PON -vs- Active Lets Get Ready to Rumble
• Available Bandwidth per subscriber
• GPON Dynamic Burstable to 2.4 Gbps
• Active Fixed 100 Mbps or 1 Gbps Often limited
  by backhaul
• Reach
• 12 Miles for GPON Active
• Can extend with remote terminals
• Deployment costs
• Active OSP costs 15-20 higher for
  enclosures,power, cooling
• Operational costs
• Much lower long-term with GPONspassive
  architecture

17
Battle of The Bandwidths

• PON -vs- Active Technical Decision?
• Full support for legacy (DS0) services
• GPON Stratum clocking GEM All services
  supported
• Active Difficult/expensive Maybe T1 Barnacle
  add-ons
• GPON benefits from statistical gain dynamics
• Unused bandwidth available to all users
• Non-video subscribers
• Inactive or low-use subscribers
• GPON better handles burst-mode traffic better
• 2.4 Gbps -vs- 100 Mbps or 1 Gbps for Active
• GPON supports more videoservice methods
• In-band IPTV
• RF overlay on 1550 nm
• Broadband (real Internet) TV

18
Battle of The Bandwidths

• PON -vs- Active Knockout Punch
• Joint Procurement Consortium (JPC)
• Verizon ATT select PON
• (BPON GPON)
• Verizon passed over 6 million homes to date
• Target is 18 million by 2011
• Largest FTTP installed base by far
• 89 of N. American FTTP is PON
• 93 of deployments since 2004 are PON
• GPON deployment rate is increasing
• More deployments higher volumes economies of
  scale decreasing component costs increasing
  savings!

Sources RVA Market Survey, 2006
19
Battle of The Bandwidths
RED HERRING ALERT!!!

• Beware of dedicated bandwidth claims
• All ports will always try to send at full rate
• Bottleneck occurs upstream in network fabric and
  uplinks

• Beware of paying for tomorrows infrastructure
  projections with too much of todays money
• Be realistic about bandwidth usage subscriber
  behavior
• Keep your service delivery options open

20
Delivering Video Services
21
Delivering Video Services

• IPTV Multicast Channel Usage Survey
• Whos watching what and how much?
• Survey conditions
• Peak use 7 PM to midnight
• During sweeps
• Service offering
• 300 broadcast channels
• 30 PPV
• 42 digital audio

22
Delivering Video Services
IPTV Multicast Channel Usage Survey
Different Channels Viewed Simultaneously
64 ONTs / PON
32 ONTs / PON
Number of Homes Served
23
Delivering Video Services

• How Much is Too Much?
• Theoretical capacity of 2.4 GPON using MPEG-4
• 1500 x 1.5 Mbps SDTV channels, or
• 350 x 7 Mbps HDTV channels
• Whats the typical channel load?
• Competitive example Dish Network
• 280 SDTV channels
• 32 HDTV channels
• Their best doesnt come close to the theoretical
  limits!

2.4 GPON
1500 SD Channels 350 HD Channels
24
Delivering Video Services
In-Band IPTV Wavelength Usage
25
Delivering Video Services

• In-Band IPTV Perspectives
• Very difficult for independent telcos to
  implement
• Large capital investment required
• Heavy churn in vendors and technologies
• DRM content licensing issues
• NRTC IP Prime similar
• Content/rights aggregators
• Turn-key service providers
• Much less capital far fewer headaches
• Probably best approach in near term
• Be careful of long-term commitment

26
Delivering Video Services
RF Overlay Wavelength Usage
27
Delivering Video Services

• RF Overlay Perspectives
• Does not compete for bandwidth with data and
  voice
• Retain existing headend if cable MSO
• Use existing set top boxes
• Cost effective
• Verizon Fios uses RF overlay
• Most flexible solution
• Wait out industry churn
• See if broadband video wins out over all others

28
Delivering Video Services

• Broadband TV The Revolution Is Being Televised
• Video programming directly from the Internet
• TV Shows Movies TiVo/Amazon, Apple, Walmart
• ABC, NBC, CBS, FOX, etc.
• YouTube, other user generated content sites
• Peer-to-peer sites Joost, Vuze, VeohTV,
  Babelgum, etc.
• A ripple now, a tidal wave tomorrow?
• Low-res rapidly giving way to HD content
• Major deals to offer mainstream series content
• Advertising supported with major money
• Strong movement toward tiered Internet
• Net neutrality in danger RBOCS want it killed!
• Charge by traffic type, i.e. The Premium Pipe
• Regulatory, business models, etc. will need to
  change

29
Getting Fiberfrom Here
To There
30
Getting Fiber From Here to There

• PON Architectural Choices
• Centralized Splitting
• Partially Distributed Splitting
• Fully Distributed Splitting

31
PON Architectures

• Centralized Splitting

Connectors (NID)
Splitters are here
Local Convergence Point (Splice)
NAP (Splice)
Central Switch
Feeder
Drops
Distribution
32
PON Architectures

• Centralized Splitting
• Target applications
• High customer churn
• Requirement for highly flexible connectivity
• Homerun consolidates all splitting to the CO
• Most flexible due to central splitting
• Highest headend flexibility/scalability
• Requires the most amount of fiber
• Most expensive, most flexible

33
PON Architectures
Partially Distributed Splitting
Connectors (NID)
Local Convergence Point (splitter)
NAP (Splice)
Central Switch
Distribution
Drops
Feeder
Splitters are here
34
PON Architectures

• Partially Distributed Splitting
• Target applications
• Overbuild with anticipated customer churn, slower
  build out, or lower-to-mid-level take rates
• Consolidates local subscribers to central
  splitter cabinet for adds drops
• Reduces feeder fiber needs
• Greater fiber usage in distribution
• Second most expensive design

35
PON Architectures
Fully Distributed Splitting
Connectors (NID)
NAPs (1xn Split)
Splitters are here
Local Convergence (1xn Split)
Central Switch
Drops
Distribution
Feeder
(Max. Take Rate)
36
PON Architectures

• Fully Distributed Splitting
• Target applications
• Higher take rates
• Lower anticipated customer churn
• Fiber lean distribution and feeder
• Least expensive up front cost
• Headend does not scale as well as previous
  architectures
• Requires higher take rates to offset investment

37
PON Architecture Summary
Cost
Flexibility
Application
Architecture Type
Higher Take Rates
Fully Distributed Splitting

Least
Low Customer Turnover
Low to Mid Take Rates/Slow Build
Partially Distributed Splitting

Mid
High Customer Turnover
High Customer Turnover

Centralized Splitting
Most
Need for High Flexibility
Cash to Burn
38
Legacy Service Considerations

• Copper-Based Services Still Required
• Low-runner legacy services
• DDS, DC alarm circuits, etc
• Regulated services such as COIN, etc.
• T1 async T1 services
• Cellular backhaul
• An ideal FTTx solution must satisfy these needs
• Add-on or barnacle equipment from multiple
  vendors
• Maintain limited copper network
• Platform that seamlessly combines copper fiber
• Fully integrated support for legacy services
• Available through the ONT via short copper loop
• Must handle all DS0-based services

39
The Optimum Optical Solution

• 2.4 GPON With Integrated Legacy Services
• Modular ONT architecture
• Provides POTS, Ethernet, and RF video interfaces
• Expandable with additional Ethernet, T1/Async T1
• HPNA V.3 and future support for other interfaces
• Provides common administration management
• Reduced training maintenance costs
• Simplified provisioning
• Less need for spares
• Supports all different split architecture choices
• Also add Really Fully Distributed option of
  putting the OLT in the remote loop carrier

40
The Optimum Optical Solution
Connectors (NID)
Really Fully Distributed Splitting
NAPs (1xn Split)
Splitters are here
Fiber Drops
Copper and Fiber Loop Carrier (1xn Split)
Central Switch
Feeder
Distribution
Copper Drops
Many OLTs share common feeder transport fibers
41
The Optimum Optical Solution

• The Ideal Integrated Platform
• 2.4 GPON for the FTTP answer
• Flexible, modular ONT architecture
• Implemented on a platform that supports all
  legacy services with integrated management
• One platform provides all this today?

42
Thank You