Session: Mo.3.F.1 Heterogeneous access fiber … access fiber networks enabled by ... UMTS GSM/EDGE...
Transcript of Session: Mo.3.F.1 Heterogeneous access fiber … access fiber networks enabled by ... UMTS GSM/EDGE...
1 ECOC 2013-09-23, Session: Mo.3.F.1
Ronald Heron – FTTH Technologies, CTO team, Alcatel-Lucent, Ottawa, Canada
September 23rd, 2013
Session: Mo.3.F.1
Heterogeneous access fiber networks enabled by multi-wavelength PONs and virtualization
2 ECOC 2013-09-23, Session: Mo.3.F.1
Abstract
• Once devoted solely to FTTH,
access fiber must increasingly support enterprises, G.FAST nodes, multi-operator uses and mobile backhaul/fronthaul for small cells. These heterogeneous networks will be supported in an operationally efficient way by multi-wavelength PONs and increasing virtualization.
3 ECOC 2013-09-23, Session: Mo.3.F.1
Fiber & Mobile Trends and Drivers
• Fiber trends:
Subs growing
BW growing …but growing slower than BW provided
Fiber Applications:
FTTH, FTTdp, Enterprise, Multi-operator, Mobile
• Mobile trends
More subs, more devices
more BW, more nodes
large cells & metro cells
CPRI BH / mobile Fronthaul
Virtualization / Compression
…fiber to the everything …fiber meets mobile
PON Aggregation market (annual sales)
EPON
GPON
10GEPON
XGPON
0
2
4
6
8
10
12
14
16
HSPA
UMTS
GSM/EDGE
GSM/GPRS
GSM
Population
Households
Average of 4.3 mobile
subscriptions per household!!!
Source:
Pyramid Research – Mar-12
Mobile growth
Source: Infonetics 2013
4 ECOC 2013-09-23, Session: Mo.3.F.1
Outline
• Fiber and mobile trends
• Fiber technology evolution and economics
• Applications
FTTH
FTTdp
Enterprise
Multi-operator
Mobile backhaul & fronthaul
• Summary and Conclusions
5 ECOC 2013-09-23, Session: Mo.3.F.1
NGPON2
Post-NGPON2
Fiber Technology Evolution in 4 Dimensions
TDM
WDM
GPON
XGPON1
Bit interleaved
40G PON
Bit internleaved
N * 40GPON
Existing 2013
TWDM GPON?
1G
TWDM
Coloured P2P – 1G, 10G
P2P
10G
2016 2020 2013 2016 2020
NGPON1
NGPON2 4-8l
Colourless WDM-PON
NGPON2 1l
Improvements in
tuning technology and
parallel optics
Improvements in
tuning technology and
parallel optics
parallel optics Improved sensitivity,
optical budget and
power consumption
6 ECOC 2013-09-23, Session: Mo.3.F.1
Fiber Technology Overview 1/3: TDM PON
• EPON, GPON Can reduce split to increase bit rate/sub
lowest cost solution and meets foreseeable BW growth
• XGPON G.987 - 10G/2.5G
can coexist with GPON
high cost, very low projected volumes
• 10GEPON IEEE 802 - 10G/1G, 10G/10G
can coexist with EPON & 10/10GEPON can coexist with GPON
high cost, low projected volumes
• NGPON2-1l-10G G.989.2: 10G/10G, 10G/2.5G, 2.5G/2.5G
can coexist with GPON, EPON & XGPON
too early to project volumes, high costs, perhaps it supplants XGPON.
GPON
GPON
GPON
GPON
G ONT 4 x 2.5G
GPON
XGPON
G ONT
XG ONT
2.5G + 10G
1580nm
1260nm
GPON
10GEPON
G ONT
10G ONT 2.5G + 10G
1580nm
1260nm
Coexistence possible with 10/10G
GPON 2.5 + 1 x 10G
NGPON2 l1
NG ONT
G ONT
1600nm
1530nm
GPON
XGPON
10GEPON
NGPON2-1l
7 ECOC 2013-09-23, Session: Mo.3.F.1
Fiber Technology Overview 2/3: TWDM-PON
• NGPON2 – 4-8l – 40G+
G.989.2: Nx10G/10G, Nx10G/2.5G, Nx2.5G/2.5G
Coexists with GPON, EPON, XGPON
Limited use cases, too early to project volumes, high costs
• GPON+ – 4-8l -10G+
Apply NGPON2 (G. 989.2) l mgmnt to GPON G.983.2 MAC
Nx2.5G/1.25G
use GPON band (other bands possible)
GPON+ could become the new GPON Can start with one lambda and evolve to more
Could it play a role in the future?
GPON
2.5G + 4 x 10G NG ONT
G ONT
NGPON2
G+ ONT 4l
GPON+
WM
4 x 2.5G
GPON+
NGPON2
8 ECOC 2013-09-23, Session: Mo.3.F.1
Fiber Technology Overview 3/3: WDM-PON & P2P
• P2P l overlay
A few coloured wavelengths are overlaid to serve a special application (e.g. mobile node)
Could be 1G or 10G
• WDM-PON
Routed OSP vs Tuned ONTs
Objective to have >32 x 1G
10G more difficult
Technical challenges
low cost tunable ONTs (Tx & Rx)
Parallel optics at OLT
• P2P
Simple P2P fiber links
1G or 10G according to the need
Technical challenge: Density - parallel optics
S ONT
32 x 1GE
x 10G WDM
PON
WDM-PON Tuned ONTs
10G
ONT
32 x 1GE
x WDM PON
10G 1GE
1GE
1GE
1GE ONT
x . . .
36 x 1GE
WDM-PON Routed OSP
GPON
2.5G + n x P2P GE GE ONT
G ONT
GE OLT GE OLT
GE ONT
P2P l overlay
P2P
9 ECOC 2013-09-23, Session: Mo.3.F.1
2.5G 10G 40G-ready
Cost Comparison
4km
•1:32 split
•Excludes civil works and OSP
construction
10 ECOC 2013-09-23, Session: Mo.3.F.1
Technical enablers
• Fiber enablers Tunable Tx for ONT
Heat-only DFB laser with cyclic AWG
TEC-based wavelength control
InP or Silicon optics tunable transmitter
Tunable Rx for ONT Tunable Bragg filter
Component for GPON+ since <2010
NGPON2 in progress
InP or Silicon-optics tunable filter
DWDM Tx/Rx for OLT
P2P, TWDM, WDM Arrays
Higher speed TDM and increased optical budgets SOA, EDFA
Coexistence elements
• Copper enablers Vectoring
Bonding
G.FAST
• Mobile enablers CPRI backhaul
Compression
Virtualization
11 ECOC 2013-09-23, Session: Mo.3.F.1
Outline
• Fiber and mobile trends
• Fiber technology evolution and economics
• Applications
FTTH
FTTdp
Enterprise
Multi-operator
Mobile backhaul & fronthaul
• Summary and Conclusions
12 ECOC 2013-09-23, Session: Mo.3.F.1
FTTH Application 1/2
• Bit rate projection
Announced service rates are increasing
Verizon vs Comcast: 100Mbs, 300, 305, 500, 505Mbps…
Chatanooga, Google fiber in KC, Gigabit Squared, fiber communities: 1Gbps
Actual continuous BW consumed is much lower (Even with aggressive assumptions)
<50Mbps by 2025
Requires ability to burst briefly at higher rates
0
200
400
600
800
1000
1200
2012 2014 2016 2018 2020 2022 2024 2026
Agg
rega
ted
de
man
d (
Mb
/s)
99%-ile aggregated streaming video demand
32 subs
16 subs
GPON is expected to be able to meet FTTH bandwidth demand far into the future
Assumes continuous increase in quality of HD, annual improvement in
compression, annual upgrade of consumer’s CPE, multiple CPE/home.
13 ECOC 2013-09-23, Session: Mo.3.F.1
FTTH Application 2/2
• Technical solutions to accommodate BW growth
GPON reduced-split
No ONT change-out
Requires splitter change at FDC, plant rework for spliced network.
Can pre-install CO splitter in cascade, then remove
XGPON or 10GEPON overlay
ONT change-out
No urgent need for residential
NGPON2
Too much BW for residential for foreseeable future
GPON+
Latent capability to turn all ONTs into higher speed
Soft provisioning of BW without truck rolls
Could be an interesting solution
WDM-PON – more technology than is needed for res
P2P – an option for short loops
GPON
GPON
GPON
GPON
G ONT 4 x 2.5G
GPON
G+ ONT 4l
GPON+
WM
4 x 2.5G
GPON+ G+ ONT GPON
Defer purchase of DWM laser
No WM Tuning capability in
ONT
14 ECOC 2013-09-23, Session: Mo.3.F.1
FTTdp 1/2
• Copper enablers
Bonding
Vectoring
• Sustained BW consumption for copper fed homes is no different from fiber fed homes
GPON rates are adequate for aggregation
Downstream 4-pair vectored
1-pair vectored
2-pair vectored
4-pair non-vectored
1-pair non-vectored
2-pair non-vectored 0
500
1000
1500
2000
2500
3000
3500
4000
16 32 64 128 256
Agg
rega
ted
dem
and
(Mb
/s)
No. of aggregated subscribers
99%-ile aggregated demand
sustained + 1G burst headroom
Year 2020
Year 2013
–Average # of streams per sub (λ): 1.25
–Acceptable HAS adaptation: 100%
–Percent of HD content is 3D 5%
–# MC channels offered: 500
–Zipf exponent -0.75
G.Fast – 1Gbps symmetrical over 200m
8 nodes with 16 ports
consume <2.5G by 2020
Expect FTTdp to be complementary to FTTH
15 ECOC 2013-09-23, Session: Mo.3.F.1
FTTdp 2/2
• Techical solutions
GPON reduced split
Shared or dedicated PONs
Completely adequate for aggregation
XGPON,10GEPON, NGPON2-1l overlay
A valid next technology for aggregating nodes
NGPON2 More bandwidth than needed
GPON+
Latent capability to turn GPON aggregation into 4-8 x speed
No truck role would be needed
Interesting option
WDM-PON - more BW than is needed
P2P – is always a backup option
GPON
GPON
GPON
GPON
G ONT 4 x 2.5G
GPON
G+ ONT 4l
GPON+
WM
4 x 2.5G
GPON+ G+ ONT GPON Defer
purchase of DWM laser
No WM
Tuning capability in
dp ONT
GPON
XGPON
G ONT
XG ONT
2.5G + 10G
1580nm
1260nm
XGPON/10GEPON
16 ECOC 2013-09-23, Session: Mo.3.F.1
Enterprise 1/2
• Characteristics
Different organizations
Often want dedicated system or dedicated l
Customers location
Clustered in business area enterprise-only solutions
…or distributed throughout residential areas integrated enterprise/residential solutions
Move BW than needed for res – dedicated BW
100G,
<=1G,
<=10G
17 ECOC 2013-09-23, Session: Mo.3.F.1
Enterprise 2/2 • Technical Solutions (for scattered enterprise/res areas)
Dedicated GPON with reduced split
Ok initially but need more (esp for 10G service)
XGPON, 10GEPON, NGPON2-1l
Dedicated enterprise-only PON
…or overlay on residential GPON
NGPON2
Only needed for highly clustered enterprise areas
GPON+
Would require all residential customer ONTs to be already GPON+
Does not address 10G service
WDM
Only needed for highly clustered enterprise areas
Must support 1G and 10G rates
P2P l overlay
Meets specific customer needs
PP
Always a fall back option – requires Fiber Distribution Cabinet
GPON
XGPON
G ONT
XG ONT
2.5G + 10G
1580nm
1260nm
XGPON/10GEPON/NGPON2-1l
GPON
2.5G + n x P2P GE GE ONT
G ONT
GE OLT GE OLT
GE ONT
P2P l overlay
18 ECOC 2013-09-23, Session: Mo.3.F.1
Multi-operator 1/4
• Characteristics
Different operators
-> dedicated system or dedicated ls
Various levels of sharing
Share in-building wiring but competing OSP infrastructure
Share fiber infrastructure
Churn means truck rolls
Share wavelengths
Differentiation but avoided truck rolls
Bit stream services
Some differentiation may be possible
Access virtualization could allow some optimizatio
19 ECOC 2013-09-23, Session: Mo.3.F.1
Multi-operator 2/4
• Technical solutions
GPON w/ fiber infrastructure sharing
Optimal for residential applications
But requires truck rolls for churn
XGPON, 10GEPON, NGPON2-1l
more than what is needed for rez
NGPON2
Allows wavelength sharing but is more BW than needed
GPON+
Allows wavelength sharing – zero touch provisioning paradigm
would be a very nice fit
WDM
Would provide wavelength sharing
Cost challenges
P2P – is used successfully where network built expressly with short loops
G+ ONT 4l
GPON+
WM
4 x 2.5G
Op A
Op B
Op C
Op D
GPON
GPON
GPON
GPON
G ONT 4 x 2.5G GPON
Op A Op B Op C Op D
GPON+
20 ECOC 2013-09-23, Session: Mo.3.F.1
COPYRIGHT © 2013 ALCATEL-LUCENT. ALL RIGHTS RESERVED. ALCATEL-LUCENT — INTERNAL PROPRIETARY — USE PURSUANT TO COMPANY INSTRUCTION
20
Multi-Operator, 3/4 Optimization Choices: Splitters in Fiber Distribution Cabinet or Hard-Wired
OLT port
OLT port
OLT port
OLT port
Splitters are
connectorized
Splitters are spliced
Connectorized pay-as-you-grow • Splitters are in flexibility cabinet
• Splitters filled up and OLT ports lit up one at a time as
subscribers are turned up
• High cost of fiber distribution cabinet
• Drives higher utilization of OLT ports
• Numerous failures related to connectors
Hard-wired Day-1 deployment • All HHP are hard-wired to splitters and corresponding
OLT ports, on day 1
• High OLT costs on Day 1
• Take rate determines OLT utilization rate
• Fiber distribution cabinet cost eliminated
• Enables zero-touch OSP – lower opex
• No connector problems
Cabinet
Some data suggests that a high proportion of field problems are related to connectors in FDCs
No cabinet
21 ECOC 2013-09-23, Session: Mo.3.F.1
Multi-operator, 4/4
• Virtualization of the access network could further enable multiple operators to share the active components while providing differentiation
OLT
Flow Table
Flow Table
FTTdp ONT
Flow Table
Flow Table
OpenFlow Controller
Edge Router
MDU ONT
Flow Table
Flow Table
RG
RG can also be an OpenFlow Device
OpenFlow Configuration
Point
Operates on a timescale of a flow.
Add, update and deletes flow entries in flow tables. Could be
integrated into OLT
Operates on a long timescale. Configuration
of ports and queues. Discovery of OF Logical
Switch capabilities,
configuration of tunnels
OF-CONFIG OpenFlow
Switch and topology discovery, capability
configuration for further study
RG
22 ECOC 2013-09-23, Session: Mo.3.F.1
Source: Bell Labs 2012
EXPLODING BANDWIDTH
VIDEO AND WIFI DRIVING FIXED BROADBAND NETWORK CONSUMPTION
By 2014:
70% of internet traffic will be
VIDEO
By 2016:
85% of wireless communications
will rely on FIXED networks
Mobile Backhaul / Fronthaul, 1/5
23 ECOC 2013-09-23, Session: Mo.3.F.1
Enterprise Small Celll
Home Small Cell
Metro Cell
Enterprise Small Cell
WiFi Hotspots
Distributed Antenna
3G/4G Macro 3G/4G Macro
Wi-Fi
A HETEROGENEOUS NETWORK PUTS NEW CHALLENGES ON THE BACKHAUL NETWORK
Mobile Backhaul / Fronthaul, 2/5 In FTTH areas and non-FTTH areas
Macro cells and metro cells
Different organizations from FTTH, different companies
-> dedicated system or dedicated l
Backhaul vs Fronthaul
24 ECOC 2013-09-23, Session: Mo.3.F.1
Need fiber-rich backhaul or very high speed multiplexed link.
Macro Cell
Or
Micro Cell
Local or Regional
Site
(BBU Hotel)
Mobile Backhaul / Fronthaul, 3/5
• Fronthaul: CPRI links from mobile radio heads to centralized Base Band Units
Consolidates the BBU digital processing in one spot
Deployment savings, smaller footprint at the radio (LightRadio)
All of the intelligence is in the large central processing cluster
possibility of virtualization and sharing of resources at different peak hours
Traffic characteristics are different from most access traffic
High sustained bandwidth = 1-10Gbps 300Mbps – 3Mbps with compression
Low latency (round trip) = <300us
24 Backhaul is much more challenging due to very high bit rates (>>100Mbps)
25 ECOC 2013-09-23, Session: Mo.3.F.1
Mobile Backhaul / Fronthaul, 4/5 • Addressing the BW and latency challenge
Bandwidth Requirement
= 1-10Gbps
DS
US
DS
US
10Gbps 2.5 or 10Gbps 1 or 10Gbps 1 or 10Gbps
Latency Requirement
= < 0.3 ms
Typical OLT for Access
•Ethernet switching, packet inspection, etc.
0.2ms – 1ms
0.2ms – 1ms
0.2ms – 1ms
0.2ms – 1ms
Fiber Access
•Speed of light (for 20km)
•Forwarding for CBR
•FEC Coding and decoding
•Ranging
0.1ms
-
0.002ms
-
0.1ms
0.05ms
0.002ms
0.2 ms
0.1ms
-
0.002ms
-
0.1ms
-
0.002ms
-
Solution: By-pass the OLT and terminate directly on BBU Hotel
Solution - if needed: Eliminate delay by setting up ranging at time of provisioning
BW is not an issue.
Latency can be addressed with ALL technologies by terminating directly on BBU Hotel
P2P l overlay, WDM-PON, P2P TDM PON / TWDM-PON
26 ECOC 2013-09-23, Session: Mo.3.F.1
Mobile Backhaul / Fronthaul, 5/5
• FH Technical solutions
XGPON, 10GEPON, NGPON2-1l
Sufficient BW for most applications, can be overlaid on GPON
small cells
mid-sized macro cells with passive distribution directly to radio head
NGPON2
Well suited for macro cells - eliminates active aggregator
Also good for distributed metro cells
GPON+ -not enough
WDM
Macro node - could serve individual radio heads without active aggregation
Scattered nodes - would require FDC in res area to separate GPON subs from mobile
P2P overlay ls over GPON
Good way to leverage FTTH network
P2P – default solution.
Good for macro cells but requires active aggregation node
Wasteful for distributed metro-cells.
OLT
BBU Hotel
l1, l2, l3, l4
splitter TWDM-PON
l1
Cell site
WDM
BBU Hotel
XGPON
XGPON l1
Cell site
GPON
GPON
XGPON
NGPON2
OLT
27 ECOC 2013-09-23, Session: Mo.3.F.1
Outline
• Fiber and mobile trends
• Fiber technology evolution and economics
• Applications
FTTH
FTTdp
Enterprise
Multi-operator
Mobile backhaul & fronthaul
• Summary and Conclusions
28 ECOC 2013-09-23, Session: Mo.3.F.1
Mobile fronthaul Residential MDU SOHO/SME Enterprise
P2P (limited)
FIBER TECHNOLOGIES FOR EACH APPLICATION
GPON GPON
XGPON / 10GEPON
P2P
WDM-PON
P2P
GPON
P2P
WDM-PON
XGPON / 10GEPON XGPON / 10GEPON XGPON / 10GEPON
NGPON2 NGPON2
Multi-Operator
P2P
GPON w. FDC
GPON+ GPON+ GPON+ GPON+
29 ECOC 2013-09-23, Session: Mo.3.F.1
Summary and conclusions
• The access network is indeed becoming more heterogeneous, with multiple applications being added to the traditional FTTH with GPON
• GPON has more than enough capacity for long term FTTH needs.
• GPON+ could in fact be very well suited for all applications except those that require a 10Gbps service rate (i.e. FTTH, FTTdp, SoHo, multi-operator)
• Mobile Fronthaul (CPRI backhaul), in particular, calls for special solutions.
• XGPON, 10GEPON, NGPON2-1l and NGPON2 could be well suited for MFH and high end enterprise solutions
• P2P will always have it’s place, though not necessarily optimal in all cases
• P2P l overlay and WDM-PON will find their niches, though work is needed to meet cost targets
• Continued innovation is needed, especially in tunable Tx/Rx and parallel optics