Ofc 2018 BBF Ng-Pon2 Workshop

Download as pdf or txt
Download as pdf or txt
You are on page 1of 296
At a glance
Powered by AI
The key takeaways are about the NG-PON2 council's organization, mission and activities.

The NG-PON2 council is organized under the Broadband Forum to drive a converged optical access network architecture and market development. Its mission is to support broadband services and drive NG-PON2 enhancements through standards.

The NG-PON2 council participated in workshops at FTTH Council Europe, OFC LA, FiberConnect Orlando, BASE BBWF Berlin, and BASE Connexion Las Vegas in 2017.

NG-PON2 workshop

Organized by Broadband Forum NG-PON2 Council

will begin at 8:30


Bernd Hesse
NGPON2 Council Chairman, BASE Chairman, Broadband Forum
Sen. Director Technology Development, Calix
[email protected]

1
Thank you to our sponsors!

2
NG-PON2 Council Overview
• General
– The NG-PON2 council is organized under the Broadband Forum
• Bi Weekly regular calls
• Sub groups (Content Group, Marketing Group, Event Group)
– Open volunteer group. Any BBF Member may join the group
– All members of the NG-PON2 council are expected to actively support the council activities
– No face-to-face meetings are required for participation
– There are no fees for joining
• Mission
– Drive a converged optical access network architecture and FTTx market development
– Support all existing and upcoming broadband service offerings
– Drive NG-PON2 enhancements {e.g. Channel Bonding , higher power Optics (e.g. N2)} and higher rate
options through the standards
– Support new applications to the technology (5G backhaul, fronthaul) as appropriate

3
Events summary
Broadband Access Summit Event BASE 2017
• Broadband Forum organized 5 workshops in 2017
✓ FTTH Council Europe (NG-PON2 Council)
✓ OFC LA (NG-PON2 Council)
✓ FiberConnect Orlando (NG-PON2 Council)
✓ BASE BBWF Berlin
✓ BASE Connexion Las Vegas

• BASE Focus areas in 2017 were:


• Optical Access Technologies (NG-PON2)
• Copper Access Technologies (GFAST)

• Workshops are educational event to update the market on the latest technology and use cases for
innovative access technologies. Workshops are industry events, to provide the audience an opportunity to
hear from industry experts on technology advantages, updates and readiness for deployments.

4
NG-PON2 Council workshop OFC, March 15th 2018

Segment Abstract Moderator

Worldwide Access Market update Lead analysts from market research Bernd Hesse
companies will provide market
updates in the optical access Chairman BASE & NG-
market segment PON2 Council
Component Market update Optical component vendors provide Ronald Heron, Director
status, challenges and innovations
on key passive and active optical Network & Portfolio
component Strategy, NOKIA
Ecosystem Overviews & System Vendors will present Kenneth Gould, Senior
Applications individual solution offerings and
capabilities to integrate NG-PON 2 Director Cable MSO,
solutions. Calix
Consultants will provide their view
on access planning scenarios
5
Market update
Segment 1

Moderator: Bernd Hesse


NGPON2 Council Chairman, BASE Chairman, Broadband Forum
Sen. Director Technology Development, Calix 6
Segment 1: Market update

8:40 – 9:05 Forces Driving and Thriving All Fiber Deployment


Heather Burnett Gold President and CEO, Fiber Broadband Association
This presentation will show the need for more fiber for 5G, video growth and IOT.

9:05 – 9:30 NG-PON2 Market and Ecosystem Update


Julie Kunstler Principal Analyst, Ovum
This presentation will review the NG-PON2 market including equipment shipments, forecasts
and prices. An overview of the end-to-end ecosystem will be provided, including optics
vendors, equipment vendors and CSPs (communications service providers).

9:30 – 9:55 NGPON2 status within Verizon


Vincent O’Byrne Director of Technology, Verizon
This talk discusses Verizon experience and lessons learned in the course of the NG-PON2 RFP,
and the plans for the NG-PON2 deployment this year.

7
Market Update
Forces Driving and Thriving All Fiber
Deployment
This presentation will show the need for more fiber
for 5G, video growth and IOT.

Heather Gold
President & CEO
Fiber Broadband Association

8
Accelerating the Connected Future

9
FTTH: Incredible Growth in North America

Sept. 2001 Sept. 2017 Sept. 2001 Sept. 2017

United States North America


34.5 Million Homes Marketed 49.3 Million Homes Marketed
15% Growth in 2017 8% Growth in 2017
15.4 Million Homes Connected 18.9 Million Homes Connected

Source: RVA, LLC

10
United States FTTH Deployment

Telephone - ILEC Tier 1 74.4%

Telephone - ILEC Tier 2 & 3 11.1%

Private Competitive - CLEC 5.6%

MSO/ Cable 3.8%

Public Municipality/ PUD 3.3%

Real Estate Development… 1.2%

Rural Electric Coop 0.5%

Source: RVA, LLC

11
Like the Internet, Fiber has Changed

12
A whole new world of
seamless communications to
EVERYWHERE AND EVERYTHING

13
An Innovation Revolution, Driven by Fiber
Applications

Networks Consumers

14
Economic Development: Fiber’s Killer App
Ten Year Gross Metropolitan
Product:
64% Better For All Fiber Cities

Ten Year Job Impact:


72% Better For All Fiber Cities

Better Year New Business


Formation:
46% Better For All Fiber Cities
Source: RVA, LLC

15
Local Experts See Fiber Impact on Local Jobs and GNP Growth

88% 85% 77%


New
52%
Employer Employer Home-based
employer
expansion retention businesses
attraction

Based on a survey of local economic policymakers,


RVA, LLC for Fiber Broadband Association, May 2017

16
Mounting Evidence Says Fiber is a Big Add
MDUs
MDU residents are willing to pay 2.8% more to purchase
a condo or apartment with access to fiber optic service.
Renters are willing to pay a premium of 8% (based on a
$1000 monthly rent) for access to fiber.

GDP
A 2014 study found higher per capita GDP (1.1%) in
communities where gigabit Internet was available. Fiber
communities enjoyed approximately $1.4 billion in
additional GDP over other similarly situated communities.

Home Values
Access to fiber may increase a home’s value by up to 3.1%.
17
All Fiber Creates Real Estate Value
Value of Amenities to Single Family Home

Source: RVA, LLC 18


All Fiber Creates Real Estate Value
Value of Amenities to MDU Home

Source: RVA, LLC


19
Now: Fiber Feeds the Innovation Economy
Smart Cities
From electrical grid management
Streaming Video to traffic management
Video accounts for 70% of
Internet traffic—and growing 5G Next Gen Services
Enhanced mobile broadband, machine
communication, ultra-reliable

Cloud Computing
Internet of Things
Computing, hosted servers,
storage, and backup need to 50 billion connected devices by 2020;
move terabytes of data, quickly hundreds in each home

The industry is driving fiber deeper and deeper into networks


to support the way we work, play and live now.
20
Cloud Computing Needs Big Bandwidth
Cloud computing
performance depends on
the internet connection –
for both the front end and
back end. More than 1
Exabyte of data is stored
in the cloud, and as
bandwidth requirements
increase, fiber will be the
only technology that can
support the necessary
network architectures.

21
Streaming Video will Continue to Explode

of North America’s consumer internet


36% video traffic grew in 2015

3.1X minutes (830,048 years) of


436 B video content will cross the
Internet each month in 2020
consumer internet video
traffic will grow at a CAGR of
in North America, percent of all
26% between 2015 and
2020 85% consumer internet traffic that will
be video by 2020

Source: Cisco VNI Complete Forecast Highlights

22
All Fiber Creates Community Value
Importance of Factors when Moving

Safe streets 98%


Very high speed Internet access 91%
Affordable housing options 90%
Greenspace/ walking, jogging, biking 85%
Reasonable commuting time 84%
Mix of housing choices 83%
Vibrant entertainment and culture 78%
Shops and restaurants walking distance 73%
Employment opportunities 71%
Affordable, convenient transportation 67%
Great school system 65%
River front w/ recreation, restaurants 54%
Professional, college sports teams 36%

Source: RVA, LLC

23
Smart Cities Need Smart Infrastructure

Smart Smart
Health Mobility City
Smart Sensor
Wi-Fi
Grid Network

Energy Efficiency Healthier Cities Civic IoT Safer Streets Connected Community
EPB in Chattanooga US Ignite and cities Verizon and the City Santa Monica City
Hiawatha Broadband in
built out a fiber around the U.S. (and of Boston are using Net provides fiber-
Minnesota piloting
network to reliably the world) are sensors and advanced supported Wi-Fi to its
project to use its fiber
manage its energy developing a smart city traffic signal controls residents in public
as a platform for home
and electrical app store predicated to measure traffic, places
monitoring of patients
systems on big bandwidth improve safety
with dementia

24
Smart Cities Need Smart Infrastructure
If you do NOT get a fiber backbone for your Senior managers in state and local
city, it may well trigger a new generation of government, survey by the Governing
economic distress. Quoting once again: Exchange:
"Fiber networks are seen by many as one of
the most important infrastructure 70% believe fiber networks should be
developments of the 21st century.” considered a public good that government
Jesse Berst: Smart City Council regulates and sometimes runs, similar to
water, sewer and other utility services.

Fiber backhaul can be a middle-mile


network within the smart city
ecosystem. High-bandwidth links
between different parts of the city can
serve as conduits for shorter cable
runs to traffic lights, municipal
buildings, surveillance cameras and
similar assets.

25
Speed, Performance of 5G Needs Fiber
Enhanced Mobile Broadband
Avg. + peak channel capacity
Ultra-reliable, Low Latency
Total network capacity Communication
Ubiquitous availability
High-reliability
High mobility, fast turnover
Guaranteed availability
Low latency
Massive Machine to
Machine Communication
Huge amount of devices
Scarce short messages
Random, connectionless
Low power, low cost 26
Stepping Stones to 5G, Paved with Fiber
Network Densification
Femto Cisco estimates small cells will have increased
11-fold between 2013 and 2018.
cell

Backhaul
Micro Pico
Small cells need expanded backhaul
cell cell capabilities. As in other places in the network,
fiber is the backhaul solution.

Metro Mobile Providers in Need


cell A report from Strategy says providers with the
largest installed base of fiber will win the day.

27
5G Fiber Needs
The ITU-T defined 5G base station requirements
to be 20 Gbps download and 10 Gbps upload.
This can only be realized through fiber-based
networks.

5G will deliver more than 10 Gbps speed, connectivity


for IoT devices, high speed mobility.

Ultra low latency is critical for self-driving connected


cars, remote robotic surgery, industrial automation
and big data transfer.

Requires massive new fiber deployment for coverage,


as shown on next slide

28
Densification Requires Much More Fiber
Going from 3G to 4G requires 25X more fiber.
Going to 5G requires at least 16X more fiber.

3G 4G 5G
1 site every 10 km 1 site every 2km 1 site for every 0.5 km
Cell density=1 cell/100 km2 Cell density= 5 x 5 Cell density= 20 x 20
= 25 cells/100 km2 = 400 cells
29
Bandwidth Deluge from IoT to be Fiber-Fed
1 Wearables
Fitness Trackers, Smartwatches, VR
5 Healthcare
At-Home Care, Remote Monitoring,
Headsets, Headphones, Health Trackers, Post-Surgery Analytics & Diagnosis,
Wearable Cameras, Smart Clothing Health Information Technology

2 Energy
Active Energy Management, Wind,
6 Manufacturing
Process Control, Maintenance,
Solar, Hydroelectric, Oil & Gas, Grid Manufacturing Execution Systems, Asset
Management Location, Smart Agriculture & Farming

3 Intelligent Shopping
Supply Chain Control, NFC Payments,
7 Security & Public Safety
Video Surveillance, Access Control,
Games, Smart Restocking & Rotation Radiation Monitoring, Other Hazards

4 Home Automation
Smart Home Management, Energy,
8 Connected Car
Fleet Management, Asset Tracking,
Water, Security, Climate Control Infotainment

30
Fiber Optic Sensing
Fiber optic sensing uses laser interrogation of
fiber optic cable to remotely and instantaneously
detect pipeline leaks, vehicle traffic, human
traffic, digging activity, seismic activity, unsafe
temperatures, loss of structural integrity, and
other conditions and activities.

On a constant basis, fiber optic sensing


accurately pinpoints the location of events from miles away.

31
What Keeps There from Being More Fiber - Average Rating: Deployment Challenges

Funding 1.7

Regulation 2.1

Technical 2.8

Marketing 3.1

Lower numbers equal greater importance

Source: RVA, LLC

32
In Rural Areas, FTTH Availability is Lowest but Share is Highest

15% 26% 30%


Availability Availability Availability

63% 43% 50%


Average Average Average
Take Rate Take Rate Take Rate

Source: RVA, LLC

33
Key Public Policy Focus in North America
Barriers to Investment Community Rural
Deployment Incentives Broadband Broadband

Encouraging forward- Pursuing light-touch Advocating for an Encouraging


leaning rights of way, regulation to benefit ownership-agnostic government support
pole attachments, consumers view of fiber networks for Universal Service
battery back-up to support facilities
policies and more to buildout
help build fiber faster

Federal, State, and Local


34
Fiber Broadband Association Educates Consumers

35
2018 Fiber Connect: Largest Fiber Conference in North America
Where can you network with the leaders in the all fiber industry?
Where can you learn how to efficiently and effectively deploy a new network – regardless of
ownership?
How do you grow and monetize your existing networks?
How will you provide the cutting-edge services key to customer satisfaction?

Meet us in Nashville to find out how!

Register online at fiberconnect.org.


36
For more, visit fiberbroadband.org.
Thank you.

37
Integration and applications update
NG-PON2 Market and Ecosystem
Update
This presentation will review the NG-PON2 market
including equipment shipments, forecasts and
prices. An overview of the end-to-end ecosystem
Julie Kunstler will be provided, including optics vendors,
Principal Analyst equipment vendors and CSPs (communications
Ovum service providers).

38
NG-PON2 Market and Ecosystem Update
Julie Kunstler, Principal Analyst, Ovum
[email protected]

15th March 2018


San Diego, CA
Where we are today – where we came from – where we are heading –
Across the ecosystem – involving component & equipment vendors & service providers

Standards – initial and amendments

Tunable optics for access - R&D Tunable optics – further R&D

More Deployments
Solution (h/w, s/w) R&D Initial Deployments
of NG-PON2 Use of NG-PON2 for “unified” access
OMCI sharing

Optics & Equipment Testing Case studies

Interoperability Testing Future amendments to standard

Trials

“yesterday” TODAY “tomorrow”

40 Ovum | TMT intelligence | informa Copyright © Informa PLC


From an analyst’s pen (computer) – selected publications

TWDM PON brings faster FTTx Tracking of NG-PON2 shipments joins our
network monetization (Oct 2014) quarterly market share reporting

Energia signals that TWDM PON has moved


to field-testing (Jun 2015) Reporting on unified access

Verizon tests NG-PON2’s wavelength and


restoration ( Sep 2015) Case studies

OFC 2016: Focus on next –gen PON


(Mar 2016)
Status of ecosystem including R&D
NG-PON2 is a key strategy for some;
expect momentum (Nov 2017)

Optical component and equipment forecasting for NG-PON2: Units, ASPs and Revenues

“yesterday” TODAY “tomorrow”


41 Ovum | TMT intelligence | informa Copyright © Informa PLC
The monetization vision was based on NG-PON2’s capabilities - 2014

▪ PON has supported non-residential applications/customers since 2006 (perhaps earlier).


▪ Consequently, monetization models were designed focusing on NG-PON2’s capabilities:
From 2014:
Revenue Scenarios using North America ARPUs (2014)
CSPs (communications service providers) want an
access solution with higher bandwidth and enough
flexibility to also support large enterprises and
fronthaul.
TWDM PON fulfills this role and provides pay-as-
you-grow options, allowing realistic adoption of
FMC (fixed–mobile convergence) throughout the
access network.
In addition TWDM PON enables efficient sharing of
the FTTx network, whether mandated by regulatory
authorities or chosen by CSPs through co-
investment partnerships.
The TWDM PON architecture facilitates FTTx
network monetization with its support of high-
revenue services over the same ODN (optical
distribution network) as residential subscribers.

42 Ovum | TMT intelligence | informa Copyright © Informa PLC


The catalysts for NG-PON2 have gained momentum since 2014

5G is here
Technical challenges:
• Delivering more than best
effort for enterprises.
FWA resurgence Bandwidth • Latency for “haul.”
(and more feasible
than ever)
demand by all
Organizational challenges:
• Wireless versus wireline
NG-PON2 teams.
meets access • Lack of holistic approach
and transport to access and transport
network costs.
growth

43 Ovum | TMT intelligence | informa Copyright © Informa PLC


FWA shows must faster payback than FTTP in selected scenarios
FWA network monetization – different scenarios – urban versus suburban,
developed country versus developing and different bandwidth offerings Factors to consider when
evaluating FWA versus FTTP:
• Is the FWA network for
residential customers only?
• What happens when other
types of customers are added?
• What are the upgrade paths as
subscribers demand more
bandwidth?
• MBH - model is highly sensitive
to MBH costs.

MBH and MFH:


• Can you use the same FTTx
network for FTTH and for
MBH/MFH for FWA?
• If you have fiber on every pole
for FWA, what else can that
Source: Ovum
fiber support?
44 Ovum | TMT intelligence | informa Copyright © Informa PLC
NG-PON2 for Mobile Backhaul (MBH) – becoming a transport solution

Challenge Potential Solution


Organizational silos at operator: Vendors must work with both sides of the operator;
• Wireless engineers handle MBH bring the two organizations together.
• Wireline engineers handle FTTx
Ovum’s MBH equipment
forecast:
Technical – can PON support MBH? • Yes, but operators want proof.
• Optical-fiber based MBH
Competitive solutions – there are well-known • Vendors must provide detailed analysis of equipment market is forecast to
solutions such as point-to-point fiber and pros/cons of the various solutions. grow to $5.3bn in 2022, from
microwave. • Vendors must solve their internal positioning of
PON for MBH.
$3.6bn in 2015.

• Optical-fiber is forecast to
There was slow uptake of small cells. • Small cell deployments are happening now. represent 59% of total MBH
equipment market in 2022.
Strategy – FTTx supports more than just FTTH. • PON becomes a piece of the transport solution.

45 Ovum | TMT intelligence | informa Copyright © Informa PLC Source: Ovum


NG-PON2 for Fronthaul – increased R&D efforts

Challenge Potential Solution


Organizational silos: • Vendors must work with both sides of the
• Wireless engineers handle Fronthaul. operator; bring the two organizations together. Ovum’s MFH equipment
• Wireline engineers handle FTTx forecast:

• Optical fiber dominates


wireless fronthaul equipment
Technical – can PON really handle MBH Issues to be solved – bandwidth, latency:
technical requirements? • NG-PON2 has the bandwidth. market due to bandwidth
• Different solutions for latency are being requirements.
developed.
• Optical fiber fronthaul
Competitive solutions – there are well- • Vendors must provide detailed analysis of equipment market is forecast to
known, proven solutions such as point-to- pros/cons of the various solutions. approach $800mn in 2022 from
point fiber. • Vendors must solve their internal positioning of $211mn in 2015.
PON for fronthaul.

46 Ovum | TMT intelligence | informa Copyright © Informa PLC


Key challenge to NG-PON2 deployments is cost, specifically optics cost

• Access is about low cost: 2.5G GPON BOSA under $8


ONT • XGS-PON under $60 and declining with BoB designs this year
optics

• Maximum of two wavelengths being deployed


• Intense R&D, cooperation between optics and equipment vendors
Interim • Operator-led interoperability testing of equipment/components

• Cost of NG-PON2 ONT optics justifies advantages (transport,


resiliency, regulatory unbundling, pay-as-you-grow)
End game • Within 2 to 2.5X of XGS-PON

47 Ovum | TMT intelligence | informa Copyright © Informa PLC


Next-gen PON versus non-next-gen PON equipment (OLTs & ONTs/ONUs)
revenue forecast – by region
Next-gen PON exceeds 78% of total PON
equipment revenues in 2023 New network builds:
• Next-gen PON has clear advantages
over 2.5G GPON or 1G EPON.
• Some new entrants are beginning
with next-gen PON on the network
side.

Existing network upgrades:


• Bandwidth demand continues to
increase.
• Use of PON for non-residential
applications and customers.

Subscriber side:
• Competition around bandwidth-
to-the home.
• Non-residential requires 10G and
sometimes more.
Source: Ovum
48 Ovum | TMT intelligence | informa Copyright © Informa PLC
NG-PON2 Equipment Forecast – strong ramp following early deployments

Leading regions:
• NA
• EMEA
• Asia & Oceania (ex China)

Minimal interest in China


currently. However, several
operators in China are evaluating
PON’s applicability to MBH and
MFH given 5G.

Source: Ovum

49 Ovum | TMT intelligence | informa Copyright © Informa PLC


Congratulations

The ecosystem has come together, enabling commercial NG-PON2 solutions.

50 Ovum | TMT intelligence | informa Copyright © Informa PLC


Copyright notice and disclaimer
The contents of this product are protected by international copyright laws, database rights and other intellectual property rights. The owner of these rights
is Informa Telecoms and Media Limited, our affiliates or other third party licensors. All product and company names and logos contained within or
appearing on this product are the trademarks, service marks or trading names of their respective owners, including Informa Telecoms and Media Limited.
This product may not be copied, reproduced, distributed or transmitted in any form or by any means without the prior permission of Informa Telecoms
and Media Limited.
Whilst reasonable efforts have been made to ensure that the information and content of this product was correct as at the date of first publication, neither
Informa Telecoms and Media Limited nor any person engaged or employed by Informa Telecoms and Media Limited accepts any liability for any errors,
omissions or other inaccuracies. Readers should independently verify any facts and figures as no liability can be accepted in this regard - readers assume
full responsibility and risk accordingly for their use of such information and content.
Any views and/or opinions expressed in this product by individual authors or contributors are their personal views and/or opinions and do not necessarily
reflect the views and/or opinions of Informa Telecoms and Media Limited.

51 Ovum | TMT intelligence | informa Copyright © Informa PLC


Integration and applications update

NG-PON2 Status within Verizon

This talk discusses Verizon experience and lessons


learned in the course of the NG-PON2 RFP, and the
plans for the NG-PON2 deployment this year.
Vincent O'Byrne
Director- Technology Group
Verizon

52
FiOS
Deployed BPON (2004) and GPON (2008)
Services include Voice, Data and Video, Business TDM and Ethernet
Symmetrical data tier of up to FiOS Gigabit Connection
FiOS > 6 m subscribers, > 17m Homes passed
Architecture is BPON 1:32, GPON 1x32 -- 1x64
Supports RF at 1550nm
MDU Strategy:
Fiber all the way to the Home/Living Unit
Small percentage of MDU use VDLS1/2 but these are being replaced with SFUs
Deploying Fiber One time for Business, Residential and Wireless (OneFiber)

Confidential and proprietary materials for authorized Verizon personnel and outside agencies only. Use, disclosure or
distribution of this material is not permitted to any unauthorized persons or third parties except by written agreement. 53
Drivers for Next Generation PON Evolution
Universal Access
Convergence of Residential and Business services and wireless backhaul/fronthaul
Increase efficiency of deployed fiber infrastructure and network element
Cost optimize access for broadband and business services
Migration of copper based services to fiber
Reduce reliance on copper plant -> Single fiber based outside plant
Reduce maintenance and operation cost
Wireless Backhaul and Fronthaul
Support for Macro cell site, small cell sites and 5G deployments

Competition
Evolution of access network is ongoing and heading to 10Gbps (DOCSIS 3.1/+, 10GEPON)

Confidential and proprietary materials for authorized Verizon personnel and outside agencies only. Use, disclosure or
distribution of this material is not permitted to anyCopyright
unauthorized Verizon 2017
persons or third parties except by written agreement. 54
Verizon Access Evolution

Confidential and proprietary materials for authorized Verizon personnel and outside agencies only. Use, disclosure or
distribution of this material is not permitted to any unauthorized persons or third parties except by written agreement. 55
Intelligent Edge Network: Architecture Vision and Scope
Objective?
Drive a lower cost infrastructure through Next Generation technologies and
network topology simplification
Create service differentiation through fast delivery of usage-based services
How?
Leverage Next Gen Technologies
Design a simplified and common network architecture across Verizon
Deliver unified domains between Access, Transport, Edge, and Core
Create OSS/SDN/NFV foundation to support more dynamic and application-based
services enabling network automation with end-to-end network management
capabilities

56
Access Strategy
Present

ONT BPON/GPON OLT

Residential
GWR ROADM

ONT GPON OLT


Incorporating Subs Mgmt (BNG) –Simplifying the Network
Business

• Intelligent Edge Network (iEN) with NGPON2

ONT NGPON2 OLT


Customer L1 L0 Next Gen
Premises
L2 Metro
OLT Transport
ONT NGPON2
Residential and Business
ONT
Access Aggregation

57
NGPON2

NGPON3(?)
10G

GPON

BPON

• Traffic continues to increase every year


• Speed Tiers Offerings
• OTT Video, IOT, IPTV increasing over time

Customer Traffic Continues to increase, suggesting > 10Gbps circa 2020


Confidential and proprietary materials for authorized Verizon personnel and outside agencies only. Use, disclosure or
distribution of this material is not permitted to any unauthorized persons or third parties except by written agreement. 58
An Example of NGPON2 TWDM Deployment
l1, l 2
20G/20G+
1 New Electronics in
CO
2 Channel Bonding , e.g.. Use of two wavelengths
at a time
TWDM OLT

λ4 New Combiner in CO

NG-PON2 λ3 NGPON 2 (10G/10G)


CO
TWDM λ2 WM
No Change in OSP
λ1 NGPON 2 (10G/10G)

Office Park
Splitter 1x32 Drop
WDM

GPON
Today 1x32 3 New NG ONT
RF video GPON + RF
NGPON 2 (10G/10G) + RF
NGPON (10G/10G)

64-128 way split


Confidential and proprietary materials for authorized Verizon personnel and outside agencies only. Use, disclosure or
distribution of this material is not permitted to anyCopyright
unauthorized Verizon 2017
persons or third parties except by written agreement. 59
NG-PON2
(PAYG-> Multiple Wavelength,
Lesson Learned Tunable Wavelengths)

Need to think “Where the Puck is


going to be”
10-15 years
Platform for 10-15 years
PAYG up to 40G/40G,
Ability to add capacity “Gracefully” or higher

Different l, class 2/3 Switching time for


Residential and Business Customers Business
on the same PON to lower costs (class 3 for Residential Service)

~10G initially (lower distances),


Ability to evolve to higher speeds Bonded, up to N x l
beyond (10G)

Presently Doing Interop (Verizon OpenOMCI


Need Vendor Interop
Specification)

Multi- Vendor on the same Fiber Different Wavelength

Need to be Competitive Ability to Grow to 40G/40G+

Confidential and proprietary materials for authorized Verizon personnel and outside agencies only. Use, disclosure or
distribution of this material is not permitted to any unauthorized persons or third parties except by written agreement. 60
Cost Benefits

Optics

Increased Splitting 128 (N2 Optics) ….. Higher?

Business and Residential on the same Saves Fibers and Splitters


Fiber (inc. increased Reliability)
No need to upgrade in a few years
Upgrading Once/Longevity
10Gbps and > 10Gbps is success based, rather that high fixed
Throughput Evolution Path costs

Competitive Advantage Competitive with DOCSIS 3.1, Symmetric, DOCSIS evolution

Support of Lower Latency services Quiet Window only on 1st Wavelength reducing latency to <
30us on additional wavelength(s)
CO Architecture Advantages
Integration of Sub. Mgmt into OLT

Total Cost of Ownership ~ GPON


Confidential and proprietary materials for authorized Verizon personnel and outside agencies only. Use, disclosure or
distribution of this material is not permitted to any unauthorized persons or third parties except by written agreement. 61
What's Still needed and Concluding Remarks
• Optics Ecosystem still needs more options to help drive costs lower
• Short Term:
– N2 Bandwidth
Tunibility
– Triplexor Equivalent (Inc. RF)
• Longer Term: Multi-Wavelength Green (CO)
Operational Benefits
– Bonded ONT (20-40G) Load Balancing
Protection
– Support for follow on after NGPON2 ….. NGPON3? PAYG Business/Residential on the
Smooth Evolution same fiber/PON/Wavelength
Lower Latency
Simpler OSP

• Helps Verizon Stay ahead of its customer's Business and Residential needs
• Overall Architecture Trialed in Tampa FL earlier this year
• Testing Business, Residential and Wireless Use cases
• Deploying in 2018

62
Thank You

63
Component Market update
will begin at 10:05

Moderator: Ronald Heron


Lead Technologist FTTH Technologies, CTO Team, Nokia, Ottawa, Canada
64
Segment 2: Component Market update (Part 1)
10:05 – 10:25 NG-PON2 Standardization and beyond
Denis Khotimsky DMTS, Verizon
This talk examines the specific complexities of NG-PON2 as a multi-wavelength PON system, analyzes the standards landscape, including the recent
and perspective advancements in NG-PON2 specifications, such as the optical ER vs Tx tradeoff, N2 power budget parameters, S/R-CP parameter
specification, ONU Tx parameter relaxation. It then proceeds with a discussion of the requirements for the next steps in higher-speed PON
standardization.

10:25 – 10:45 NG-PON2 Optics and Photonic Integration Approaches


Hal Roberts System Engineer and Architect, Calix
The first wave of NG-PON2 optics have been based on conventional bulk free-space optical modules. Photonic Integrated Circuits have always been
considered a promising method of producing lower cost NG-PON2 optics, however the upfront development costs and timeframe have prevented these
from being realized during early optics development. Recently promising work has begun on PIC based NG-PON2 optics. This presentation will examine
how PIC based NG-PON2 optics might accelerate the availability of standards compliant lower cost optics.

10:45 – 11:05NG-PON2 ONU w / pluggable CATV overlay solution


Ben Hur Senior Managing Director, Lightron
Based on the experience of developing 10G / 10G NG-PON2 ONT, NG-PON2 technology can be evolved into various applications by introducing the
development of NG-PON2 triplexer product with CATV overlay function/solution in the form of XFP Pluggable Optic. The NG-PON2 triplexer developed
product of this announcement provides compatibility to various platforms of NG-PON2 and provides the best solution for various additional application
fields.

11:25 – 11:45 Access finally has wavelength flexibility


António Teixeira co-founder/prof, Picadvanced/UA
This presentation will focus on technical challenges to be overcome and some compromises (and have) to be made to foster broad market adoption to
happen sooner than later. PICadvanced, in this work, will present the current status of its solutions and the results of its production units which, given
the facts, that is based on DML and fast thermal control in both Tx and Rx, show great potential to open the road to broad adoption of the standard.

.
65
Segment 2: Component Market update (Part 2)

11:45 – 12:05 Optical transceivers for NG-PON2 and beyond: an update


Wei-Ping Huang Founder and Chief Scientist, Hisense Broadband Multimedia Technologies Co.,Ltd.
We review the recent progress made in the optical transceivers for NG-PON2. The key issues to improve the cost–performance ratio are
addressed and the potential solutions are proposed.

12:05 – 12:25 Low-Cost Coherent Detection for NG-PON2


Jesper Bevensee Jensen CTO & Founder, Bifrost Communications
We present the most recent update on our effort to develop Quasi-Coherent Receiver Technologies that pave the way towards a realization of
the higher NG-PON2 Classes. THis update will include the first tests of the Quasi-Coherent receiver in combination with a commercially avaiable
ONU transmitter.

12:25 – 12:45 Discussion on low cost NGPON2 ONU solution based on DML
Dr. Ben Chen, Technical Director, Accelink Technologies
In this topic, we will like to investigate the possibility to use DML for NGPON2 ONU solution under G.989.2 Amendment 2, in order to achieve
low cost and to be ready for massive deployment.

66
Component market update
NG-PON2 Standardization and beyond
This talk examines the specific complexities of NG-PON2 as
a multi-wavelength PON system, analyzes the standards
landscape, including the recent and perspective
advancements in NG-PON2 specifications, such as the
optical ER vs Tx tradeoff, N2 power budget parameters,
S/R-CP parameter specification, ONU Tx parameter
relaxation. It then proceeds with a discussion of the
Denis A. Khotimsky requirements for the next steps in higher-speed PON
Distinguished Member of standardization.
Technical Staff, Verizon .

67
NG-PON2 Standardization
and Beyond

Denis A. Khotimsky
Verizon Network Planning
This talk is largely concerned with PON standards…

1. Common recipe for PON “alphabet soup”


2. Glimpse of NG-PON2 history
3. PMD and TC layers
4. Inter-Channel-Termination Protocol
5. Verizon Open Specifications
6. Higher-speed PON systems

69
TDM/TDMA to TWDM system evolution

Conventional TDM/TDMA PON system

ld
ONU
lu

ld
ONU
ld lu

lu ld
ONU
lu

ld
ONU
lu

Power
splitter 70
TDM/TDMA to TWDM system evolution

Multiple OLT CTs, tunable ONU transceivers

ld1 ld1, ld2, ld3, ld4


OLT CT ONU
lu1 lu1 /lu2 /lu3 /lu4

ld2 ld1, ld2, ld3, ld4


OLT CT ONU
lu2 lu1 /lu2 /lu3 /lu4
ld1, ld2, ld3, ld4
lu1, lu2, lu3, lu4
ld3 ld1, ld2, ld3, ld4
OLT CT ONU
lu3 lu1 /lu2 /lu3 /lu4

ld4 ld1, ld2, ld3, ld4


OLT CT ONU
lu4 lu1 /lu2 /lu3 /lu4

Wavelength Power
multiplexor splitter 71
NG-PON2: A Taste of the Good Life
• Non-service-affecting activation
• Possibility to eliminate service impact upon ONT activation and re-activation (no
destructive quite windows on the chosen critical wavelength channels)

• Non-service-affecting maintenance
• Objective: no maintenance-caused service outage
• Requirement: bound the service outage experienced by an in-service ONU as a result of
any scheduled maintenance operation by 50 ms.

• Protection across PON card


• Possibility to restore service in case of card failure in 50 ms

• Bonding
• Possibility to increase the maximum service rate available on the PON through aggregation
of wavelength channels without upgrading the system itself

• Load balancing
• Optimal in some sense repacking of services per wavelength channel

• Power savings
• Possibility to turn off OLT channel terminations in low load
72
Verizon Access Network Evolution 2018

40G NG-PON2
Up to
10 Gbps*/Sub
40G*/PON
Leapfrogged
• Enabled by Novel Tunable Lasers
10G • Evolution to 80G* in future
XG-PON1
• Channel Bonding for future evolution
XGS-PON
Up to
10G* per PON/Sub
GPON
1G Deployed for
< 1 Gbps
2004
50M 2008
BPON
Limited to
< 100 Mbps

100M 2.5G 10G 40G


Downstream
* These are PON line rate. Available bandwidth for customer traffic is lower because of protocol overheads
73
A glimpse of history

The term NG-PON2 has been is


use since at least 2007!

• Long before the system took shape

• For half of the time its semantics has differed from what we know
today.

74
2007 FSAN view (from Middletown meeting report)

• The NG-PON study includes the following two varieties.


NG-PON1 NG-PON coexisting on same ODN as GPON based on G.984.5 approach
NG-PON2 “disruptive” NG-PON with no requirement to coexist on same fibers as GPON

75
75
Evolution of the concept

Original concept prevailed while NG-PON1 systems were specified:


o XG-PON1 – Asymmetric 10/2.5G TDM/TDMA – ITU-T series G.987 (2010)
o XG-PON2 – Symmetric 10/10G TDM/TDMA was shelved

Specific NG-PON2 study:


o Project initialized in the Fall of 2010
o Operator requirements White Paper: April 2011
o Vendor technology White Paper:
• First CFC – June 2011
• Version 1.0 – April 2012

76
NG-PON2 Technology selection

Vendor Technology White Paper (June 2011 – April 2012)


– 40G TDM PON
– TWDM PON
– WDM PON
– OFDM PON

Quoting the FSAN meeting report from April 2012:


Operators reviewed the remaining in-scope solutions and agreed to
include the following in any future studies:
– TWDM-PON as a primary NG-PON2 solution

– With support for (tunable) WDM overlay if needed

77
NG-PON2 Standardization enablers

ITU-T: BBF:
– Requirements – Inter-Channel Termination protocol
– Physical media dependent (PMD) layer – YANG models for ITU-T PON
specification
– Test specifications
– Transmission Convergence (TC) Layer
Specification
– OMCI for NG-PON2

78
NG-PON2: Original ITU-T standardization timeline

Recommendation Content Consent Approval

G.989.1 General requirements 09/2012 03/2013

Physical media dependent (PMD) layer


G.989.2 12/2013 12/2014
specification

Transmission Convergence (TC) Layer


G.989.3 07/2015 10/2015
Specification

G.989 Definitions, abbreviations, acronyms 07/2015 10/2015

79
NG-PON2 reference architecture

80
G.989.2: PMD layer specification
Scope
▪ Wavelength plan: central wavelengths, spacing, spectral excursion
▪ Nominal line rate, optical path loss, fiber distance, tuning time classes
▪ Upstream and downstream PMD parameter tables for 2.5 and 10G IF (S/R-CG)

Multichannel tunable aspects of the specification


▪ Optical power requirements across all wavelength channels (Types A ,B)
▪ Multiple sources of spectral excursion
▪ In-band (interferometric) and inter-channel (non-coherent) crosstalk
▪ Out-of-band, out-of-channel, when-not-enabled power density specs

81
G.989.2 Amendment 1

Consented: 02/2016
Approved: 04/2016
Key developments:
▪ Completion of 8 channel specifications
▪ 0.5dB relaxation of the OPP and OLT Rx sensitivities
▪ 0.9dB relaxation of WNE and OOC PSD specification for the ONU

82
G.989.2 Amendment 2

Consented: 06/2017
Approved: 08/2017
Key developments:
ER 1

Extinction ratio
▪ ER vs Tx_min tradeoff
ER 2
▪ Extra OPP compensation with Tx_min ER 3

▪ X/S mask refinement ER 4,5

▪ OLT crosstalk tolerance refinement


– – – – –
▪ Relaxation of the ONU PLO time allocation P1 P2 P3 P4 P5
Mean launch optical power

(𝑃ത1 , 𝐸𝑅1 ) (𝑃ത2 , 𝐸𝑅2 ) (𝑃ത3 , 𝐸𝑅3 ) (𝑃ത4 , 𝐸𝑅4 ) (𝑃ത5 , 𝐸𝑅5 )
3.0 4.0 5.0 7.0 9.0
8.0 6.0 5.0 4.0 4.0

83
G.989.2 work in progress

Items under discussion (towards Amendment 3):


▪ Optical parameter specification at the S/R-CP reference point (OLT transceiver interface)
▪ Parameter relaxation through segregation of Type A and Type B upstream optical links
0.9dB
▪ Statistical (rather than worst-case) approach to parameter modeling

84
G.989.3: TC layer specification
Scope
▪ Protocol layers and functionality
▪ Data formats
▪ Physical layer operation and management functions, messages, and protocols
▪ ONU activation cycle
▪ Security
▪ Performance monitoring

Multichannel tunable aspects of the specification


▪ ONU activation in a multichannel system
▪ Wavelength mobility (controlled wavelength handover)
▪ Wavelength protection switching

85
G.989.3 Amendment 1

Consented: 09/2016
Approved: 11/2016
Key developments:
▪ Refinement of Type B (dual OLT) protection
▪ Clarification on Channel Partition Index
▪ Provisions for delay-sensitive services (cooperative DBA, increased number of bursts per
frame)

86
G.989.3 work in progress

Items under discussion (towards Amendment 2):


▪ Wavelength channel bonding
▪ PLOAM channel enhancements
▪ Clarifications and improvements for Wavelength mobility robustness

87
TR-352: Inter-channel-termination protocol
SNI

Approved: March 2017 OLT Chassis

Line card
ICTP

OLT and OLT interoperability functional


OLT
primitives CT 1

• First introduced in G.989.3; developed in detail by BBF OLT


CT 2
• Covers interactions between OLT channel terminations
➢ Within a card Line card

➢ Within a chassis OLT


CT k

➢ Across chassis, including those of different vendors.

NG-PON2

• Supports the functionality of
OLT Chassis
➢ ONT activation, authentication and provisioning
Line card
➢ Service information exchange between OLT CTs OLT
➢ ONT mobility management CT m

Work in progress


OLT Chassis
• Clarifications on the assigned numbers Line card
OLT
• Clarifications on Wavelength mobility manager CT n

88
Verizon Open Specifications

Verizon OpenOMCI
• Version 1.0 has been published (July 2017)
http://www.verizon.com/about/techspecs/verizon-openomci-specification
• Updated version being prepared
• Core MEs associated with modeling and management of the multi-
wavelength aspect of the system standardized via G.988
Verizon Open NETCONF/YANG
• Advanced management features of dual managed ONU/NIDs
• On-going development through a Verizon-led vendor NG-PON2
interoperability group

More details in the Interoperability talk…

89
Standardization of Higher-Speed PON systems

• Earlier this year ITU-T initiated several new Higher-Speed PON projects
-- G.hsp.req – System requirements
-- G.hsp.comTC – Common TC layer
-- G.hsp.50Gpmd – PMD layer for single channel 50G TDM system
-- G.hsp.TWDMpmd – PMD layer for a multi-channel tunable TWDM system

90
Verizon position on HSP standardization

• An Higher Speed PON system being specified and to be deployed around 2022 should aim at
sustaining bandwidth demand growth for the subsequent ten years (that is, the overall
capacity of several hundreds Gigabit per second)

• An HSP specification should focus on what is needed, rather than what can be specified

• The new HSP system should be cognizant of legacy deployment and use these as building
blocks

• The tremendous operational advantages of NG-PON2 should be carried forward -- Non-


service-affecting activation and maintenance
-- Line card protection
-- Wavelength mobility (load balancing, rogue mitigation, OLT power savings)
-- Higher service rate through channel bonding

• Build off NG-PON2

91
Thank you.
Component market update
NG-PON2 Optics and Photonic
Integration Approaches
The first wave of NG-PON2 optics have been based on
conventional bulk free-space optical modules. Photonic
Integrated Circuits have always been considered a
promising method of producing lower cost NG-PON2 optics,
however the upfront development costs and timeframe
have prevented these from being realized during early
Hal Roberts optics development. Recently promising work has begun on
System Engineer and PIC based NG-PON2 optics. This presentation will examine
Architect how PIC based NG-PON2 optics might accelerate the
Calix availability of standards compliant lower cost optics.

93
NG-PON2 Optics and
Photonic Integration
Approaches
Agenda

▪ Examine TWDM-PON free space optics


(FSO) approaches
▪ Comparing FSOs with PICs
▪ Photonic Materials – InP and Silicon
▪ Monolithic vs Hybrid PICs
▪ Example Transceiver – Monolithic vs
Hybrid
▪ Summary
Broadband Forum BASE Event at OFC 2018 March 15, 2018 95
Free Space Optics – TWDM-PON ONU
TWDM ONU Transceiver Components
• Detector: generally APD – Avalanche Photodiode
• Blocking Filter to allow coexistence

• Tunable Laser
• Directly modulated (DML) or Externally modulated (EML)
• EML may have Semiconductor Optical Amplifier (SOA) to provide
adequate transmit power APD
• Diplex Filter Tunable Filter
Blocking Filter
• To separate upstream λ
from downstream λ +
SOA EM Tunable Laser
Fiber Pigtail or Connector
Diplex Filter
-
Broadband Forum BASE Event at OFC 2018 March 15, 2018 96
What is a PIC?
Photonic Integrated Circuit (also called ‘Integrated Optics’)
▪ A PIC contains two or more optical components on a single substrate
▪ Light is guided by waveguides on the substrate, rather than through free
space with lenses
▪ The Externally Modulated Laser (EML) is a very simple example of a PIC
▪ A more complex example would be a full transceiver on a single substrate

Courtesy PICadvanced

Broadband Forum BASE Event at OFC 2018 March 15, 2018 97


Free Space vs. Integrated Optics?
This? Or This?

General Rule of Thumb


▪ The larger the number of individual optical components…
▪ …the more PICs become attractive
Why?
▪ Each optical component in a free space chain must be aligned to each other
▪ Lenses required to collimate and focus light along chain
▪ Optical losses accumulate
▪ Assembly becomes larger and may require larger package (XFP vs SFP)
▪ Leads to high manufacturing cost and potential fallout
Disadvantage to PICs
▪ Large upfront cost, long development time and requires high volume to justify 98

Broadband Forum BASE Event at OFC 2018 March 15, 2018


Photonic Materials for Optical Transceivers
Indium Phosphide Cons
▪ More expensive per mm2
▪ Not compatible with CMOS fabs
Indium Phosphide Pros
▪ Can generate or amplify light
▪ Potential for fully integrated optical circuit
Silicon Cons
▪ Cannot generate or amplify light
▪ Cannot be used alone for PIC transceiver
Silicon Pros
▪ Less expensive per mm2
▪ Compatible with CMOS fabs Source: http://www.jeppix.eu/multiprojectwafers-1/
Courtesy PICadvanced

Broadband Forum BASE Event at OFC 2018 March 15, 2018 99


Types of Photonic Integration
Monolithic i.e. InP Hybrid i.e. InP + Silicon
Capability of integrating multiple components Combining optical devices from different
including passive and active in a single chip materials on a common substrate
using only one substrate material

PROS CONS PROS CONS


+ Integration of + Difficult to combine + Taking advantage + Fiber coupling
active and passive electronic IC and of both material losses
devices in a photonic functions systems + Manufacturing
generic process; + InP wafers more costs of adding
+ Benefits in terms of expensive and and aligning InP
reliability, power smaller than components to Si
efficiency, and Silicon substrate
testing; counterparts

Courtesy PICadvanced and Aeponyx

Broadband Forum BASE Event at OFC 2018 March 15, 2018 100
Example: Monolithic InP Transmitter PIC
Example: SOA boosted Tunable EML Laser
▪ Red boxes are active elements on InP

+
SOA EM Tunable Laser
Fiber Pigtail or Connector
-

101

Broadband Forum BASE Event at OFC 2018 March 15, 2018


Example: Monolithic InP PIC
Example: SOA boosted Tunable EML Laser with Quasi-Coherent
Tunable Receiver
▪ Red boxes are active elements on InP

PD Tunable LO

+
SOA EM Tunable Laser
Fiber Pigtail Bragg Filter
-
102

Broadband Forum BASE Event at OFC 2018 March 15, 2018


Example Hybrid Silicon + InP PIC
Example: SOA boosted Tunable EML Laser with Quasi-Coherent
Tunable Receiver

Yellow-box is Silicon serving as packaging substrate for InP red elements with waveguides in
Silicon

PD Tunable LO

PD
PB +
SOA EM Tunable Laser
Fiber Pigtail Bragg Filter
-

Broadband Forum BASE Event at OFC 2018 March 15, 2018 103
Example Hybrid Silicon + InP PIC with MEMS
Example: SOA boosted Tunable EML Laser with MEMS Tunable
Receiver
Yellow-box is Silicon serving as packaging substrate for InP limited to red boxes
Green MEMS filter is Silicon
MEMS Tunable Filter PD
Diplexer is Silicon

+
SOA EM Tunable Laser
Fiber Pigtail Bragg Filter
-

Broadband Forum BASE Event at OFC 2018 March 15, 2018 104
Summary
Photonic Integration may lower costs for complex TWDM optics
Two major approaches are being pursued:
▪ Monolithic InP – Full transceiver possible on a single chip, but chip real
estate must be reasonably small to be cost effective
▪ Hybrid Silicon + InP – Low costs of Si substrate must be balanced against
the costs of adding and aligning active InP elements

Broadband Forum BASE Event at OFC 2018 March 15, 2018 105
Broadband Forum BASE Event at OFC 2018 March 15, 2018 106
Component market update
NG-PON2 ONU w / pluggable CATV
overlay solution
Based on the experience of developing 10G / 10G NG-PON2
ONT, NG-PON2 technology can be evolved into various
applications by introducing the development of NG-PON2
triplexer product with CATV overlay function/solution in the
form of XFP Pluggable Optic. The NG-PON2 triplexer
Ben Hur developed product of this announcement provides
Senior Managing Director, compatibility to various platforms of NG-PON2 and
Lightron Inc. provides the best solution for various additional application
fields.

107
Doc. # : R51-180219-AC1
Leveraging Essence

New Approach : NG-PON2 ONU w/ pluggable


CATV overlay solution
Ben Hur
Senior Managing Director @ LIGHTRON Inc.
[email protected]
Agenda

▪ Up-to-date NG-PON2 “ONU Diplexer” Activities


▪ Why NG-PON2 ONU w/ RF Overlay??
▪ Initial Implementations
▪ In-house Feedbacks
▪ Summary & What to do next

Page ▪ 109
All Rights Reserved. Copyright©2018 Lightron
NG-PON2 Activities :
✓ Key Enablers to a Wide Scale Adoption : Cost, Form-factors & Productivity
✓ TEC Controlled DML & Etalon-filtered APD in lieu of EML+SOA & APD

Productivity
TAM for High
Volume Product
Lifecycle Performance
Reasonable
CAPEX
$ ÷ Gbit/s & OPEX

Page ▪ 110
All Rights Reserved. Copyright©2018 Lightron
NG-PON2 Activities : Core Design Technology

✓ Transmitter TO-Can ✓Receiver TO-Can ✓4ch. Tunable BOSA Assembly


• DML LD chip (DWDM grade) • Filtered APD • Built-in Isolator
• Cooled by TEC • Cooled by TEC • 8o angled ceramic stub w/ ferrule
• 45o Mirror • Micro Lens

Cooled LD TO can with


Aspherical lens
Cooled PD TO
can.
APD chip Micro-Lens
TIA
TO Stem Thermistor

Page ▪ 111
All Rights Reserved. Copyright©2018 Lightron
NG-PON2 Activities : SFP+ & XFP Diplexers
✓ Block Diagram ✓ Product Physical Variations
MOD disable MOD_disable ▪ Both SFP+ & XFP Package Design
Combo Burst Enable Burst Enable
IC Tx-Fault
Rx_SD
Tx-Fault
Rx_SD ▪ SFP+ for Enterprise CPEs, 5G Small Cells
Tx_SD Tx_SD
Tx_Data +/-
Rx_Data +/-
Tx_Data +/-
Rx_Data +/-
▪ Various Optical I/F: LC/UPC, LC/APC
SCL
SDA MCU SC/UPC, SC/APC
2-Wire DDM I2C BUS
Interface & SDA/S I2C_SDA
SDA/M SCL/S
SCL/M
Data I2C_SCL
BOSA Process
Tx TEC Driver &
LD w/ TEC ; Tx wavelength control
LUT
DAC control
Ferrule WDM DAC0
DAC1
DAC2
Rx TEC Driver DAC3 Memory
Tunable APD-TIA ; Tx wavelength control Map
filter w/ TEC
SFP+
APD Bias Driver or XFP
; APD Bias control
Connector

Page ▪ 112
All Rights Reserved. Copyright©2018 Lightron
NG-PON2 Activities: Beef-ups & Compliances

2017 2018
SFP+Diplexer
Module Form-factor Offering SFP+ Diplexer XFP Diplexer
XFP Triplexer (in the pipeline)
Tx Power +7 ~ +9dBm
Gating Issue at PMD Tx Power +4 ~ +9dBm
ER : 4dB
(Transmit Power & ER) ER : 6dB
(Amendment 2 Executed)

Gating Issue at PMD OOC1 : < -40.5dBm/15GHz OOC1 : Comply


(OOC-PSD) OOC2 : < -44.8dBm/15GHz OOC2 : New Amendment in Progress

Gating Issue at PMD


0.32nm @ Pmax -15dB 0.2nm @ Pmax -15dB
(Burst Mode Drift, MSE)

System Level Interoperation Unready Sequence & Timing Sync-up

Page ▪ 113
All Rights Reserved. Copyright©2018 Lightron
Why “RF Overlay” is in contemplation?
✓ Interim needs prior to All-IP Services at both B/G-PON franchise area and NG-PON2 greenfields
✓ Regulations on how OTT service is done.

Legacy PON
G-PON ONT
at Brownfield

Power splitter XG-PON1


NG-PON2 ONT

NG-PON2
ONT1
NG-PON2
ONT2
NG-PON2
ONT3
NG-PON2
ONT4

1555+/-5nm

Page ▪ 114
All Rights Reserved. Copyright©2018 Lightron
Product “Target” Specifications
Parameter Min. Typ. Max. UoM Remarks
Operating Temperature 0 - 70 ᵒC
Power Dissipation - - 3 W
Digital Transmitter
Mean Output Power +7 - +9 dBm G989.2 Amendment2
Frequency 195.3 - 195.6 THz Compliant with G989.2
Tuning Speed - - 1 sec Compliant with G989.2 Class 3
Extinction Ratio 4 - - dB G989.2 Amendment2
SMSR 30 - - dB Compliant with G989.2
Digital Receiver
Sensitivity - - -28 dBm Compliant with G989.2
Frequency 187.5 - 187.8 THz Compliant with G989.2
Tuning Speed - - 1 sec Compliant with G989.2
Overload -7 - - dBm Compliant with G989.2
Parameter Min. Typ. Max. UoM Remarks
Analog Receiver
Operating Wavelength 1550 - 1560 nm
Responsivity 0.8 0.9 - A/W
Operating Frequency 54 - 1002 MHz
RF Output Level 17 - 20 dBmV 1002MHz CW signal, OMI 3.5%
RF Frequency Response Flatness - - 2 dB 54~1002MHz
RF Output Tilt 0 1.5 3 dB 54~1002MHz
RF Return Loss 14 - - dB 50ohm, 54~1002MHz
CSO 60 - - dBc
CTB 60 - - dBc
CNR 44 - - dBc
MER 35 - - dBc
Crosstalk - digital signal to analog signal - - 3 dB Added noise by digital signals.
Page ▪ 115
All Rights Reserved. Copyright©2018 Lightron
New Approach : Key Specs, Strengths & Usage on ONTs
✓Key Spec ✓Product Design Strengths
▪ Equivalent Digital Performances ▪ Robust & Miniature Optical Engine Integration
▪ Standard XFP Form-factor ▪ Excellent Thermal Management
▪ SC/APC Optical Interface ▪ Superior Analog Performances
▪ Power Supply : +3.3/+5.0V ▪ Optimized Anti-crosstalk Design
▪ Analog Receiver at 1555nm+/-5nm window
▪ 54 to 1002MHz downstream RF frequency
▪ Optical input power range (-6~0dBm  -8 ~ +2dBm)
▪ Crosstalk by PON & RF Path : < 3dB (added noise by digital signal)
▪ 50 ohm Single RF Output Interface

RF Output

Diplexer Filter
MoCA 2.0 LAN
RGMII
SMB Connector

50Ω to 75Ω Balun

PON MAC XFP Triplexer Unit on ONT Blade Optical In/Output


(SC/APC receptacle)

Recommend RF cable interface with RF jack(socket)


Page ▪ 116
All Rights Reserved. Copyright©2018 Lightron
New Approach: Product Block Diagram
Rx
Limiting Amplifier
Data I/O

Tx
Laser Driver Data I/O

TEC Driver MCU


TX

TEC Driver I2C I/O


RX
DAC
Bias Driver
Digital RX

Bias Driver
Analog RX Memory Map
Power Supply
Digital +3.3V
RF
amp Analog +5.0V
RF
Transformer Attenuator amp

RF
Analog Receiver Part amp

Page ▪ 117
All Rights Reserved. Copyright©2018 Lightron
New Approach: Pinning & High Level Components
# Symbol Description
1 GND Module Ground
2 Tx_Fault Module TX Fault
3 NC No Connect
4 Tx_SD Signal Detection when Laser is on
5 Tx disable Transmitter Disable
6 VCC5 +5.0V Power Supply
7 GND Module Ground
8 VCC3 +3.3V Power Supply
9 VCC3 +3.3V Power Supply
10 SCL 2-Wire Serial Interface Clock
11 SDA 2-Wire Serial Interface Data Line
12 Mod_ABS MOD_Absent Triplexer OSA Assembly PCBA TOP & BOTTOM
13 NC No Connect
14 RX_LOS Receiver Loss of Signal Indicator
15 GND Module Ground
16 GND Module Ground
17 RD- AC Coupled
18 RD+ AC Coupled
19 GND Module Ground
20 NC No Connect
21 MOD_DIS Module disable (Power down)
22 NC No Connect
23 RF GND RF Module Ground
24 RF Out RF Output
25 RF GND RF Module Ground
26 GND Module Ground
27 GND Module Ground
28 TD- AC Coupled
29 TD+ AC Coupled
30 GND Module Ground

Integrated XFP
Page ▪ 118
All Rights Reserved. Copyright©2018 Lightron
New Approach: Optical Ass’y Performances
Digital U/S Analog D/S Digital D/S
0.00

-10.00

Transmitance
✓ Smaller -20.00

✓ Compact design -30.00

-40.00

-50.00

-60.00
1,520 1,530 1,540 1,550 1,560 1,570 1,580 1,590 1,600 1,610
Triplexer OSA Assembly Wavelength [nm]

Signal I/O Paths Filter Spectrum

Triplexer OSA Test Data


Tx (+25°C) Rx (+25°C) RF Rx (+25°C)
Sensitivity
No. Wavelength Po MPD SE TE Vbr Vapd Overload DCR
BER=10E-3
nm dBm mA mW/mA dB V V dBm dBm mW/mA
1ch(1532.68) 7.80 0.71 0.088 0.00
1 28.63 26.43 -30.7 >-6 1.01
4ch(1535.04 7.64 0.53 0.092 1.39
1ch(1532.68) 8.06 0.71 0.098 0.00
2 27.05 26.61 -30.2 >-6 1.02
4ch(1535.04 7.71 0.53 0.102 1.30
1ch(1532.68) 8.14 0.55 0.100 0.00
3 28.55 26.29 -30.5 >-6 1.01
4ch(1535.04 7.51 0.39 0.091 0.38

Page ▪ 119
All Rights Reserved. Copyright©2018 Lightron
New Approach: Digital Transmitter Performances
✓ Eye Diagram ✓ Burst Mode Operation ✓ LD On/off Time @ Burst Modes

500ps

~ 2.xns

64us ON 64us OFF 64us ON

500ps
▪ Test Condition ▪ Test Condition
• 9.95328Gb/s, 231-1 PRBS
• Burst pattern : 64us on/ 64us off
• Filtered Eye Diagram
• Preamble pattern : 1010, 1000bit
• CH1(1532.68nm)
• Data field pattern : PRBS 231-1
• Output power : +7.5dBm ~ 2.xns
• Extinction ration : 4.15dB

✓ Burst enable/disable function is


working properly, within 12.5ns.
Page ▪ 120
All Rights Reserved. Copyright©2018 Lightron
New Approach: Digital Receiver Performances
✓ Rx Electrical Output Diagram ✓ Downstream Sensitivity
✓ max. -28dBm Rx sensitivity spec. for Type
A/N1 Class

✓ At FEC-OFF condition,
-21 ~ -7dBm sensitivity are measured(@
BER 1x10-12BER).

▪ Test Condition ▪ Test Condition ▪ Source Condition


• 9.95328Gb/s, 231-1 PRBS • 9.95328Gb/s, 231-1 PRBS
• CH1(1596.34nm) • BER 1x10-3
• Source ER : 8.2dB • CH1(1596.34nm)

• Source ER : 8.2dB

Page ▪ 121
All Rights Reserved. Copyright©2018 Lightron
New Approach: Analog Receiver Performances (RF Output)
Analog Output Waveform @ 54~1002MHz
CATV Analog
& QAM
CATV Optical
Multiple
Transmitter
Signal
Generator
Optical
PWR Meter
Analog CW 78ch
+ 256QAM 77ch(-6dB) Fiber
+ 1GHz CW 1-tone Spool
20km

Optical
ATT
Analog 20dBmV/ch
Optical Input : -3dBm QAM 14dBmV/ch
QAM &
NG-PON2 RF
: Optical signal Spectrum
Triplexer ATT
: Electrical signal Analyzer
: Measuring Optical PWR
CATV Analog & QAM Characteristic Measurement Setup

Page ▪ 122
All Rights Reserved. Copyright©2018 Lightron
New Approach: Analog Receiver Performances (CSO & CTB)
✓ CSO & CTB are dependent on X-Talks  need to secure a bit of additional margin

CSO (Composite Second Order, > 60dBc) CTB (Composite Triple Beat, > 60dBc)

CSO CTB
=62.2dBc =61.6dBc

Page ▪ 123
All Rights Reserved. Copyright©2018 Lightron
New Approach: Analog Performances (CNR & MER)
▪ CNR & MER are dependent on X-Talks  need crosstalk improvement in advance

Analog D/S CNR (>44dBc) Analog D/S QAM MER (>35dBc)

Corrected
CNR=
40.3dBc
MER=
26.6dBc

Page ▪ 124
All Rights Reserved. Copyright©2018 Lightron
New Approach: Analog Performances (X-Talk)
Analog CW 78ch CATV Analog &
QAM Multiple CATV Optical Crosstalk – Digital signal to Analog signal
+ 256QAM 77ch(-6dB)
Signal Transmitter
+ 1GHz CW 1-tone
Generator
Optical
PWR Meter
Digital  RF port’s noise level
PPG Transmitter
Data +/- before PON path On

BERT(2 ch.) Fiber


Spool
20km

▪ Test Condition 17dB X-talk


• 9.95328Gb/s, 231-1PRBS Optical
ATT
• CH1(1596.34nm) Analog 20dBmV/ch
• Source ER : 4.0dB QAM 14dBmV/ch

Data +/- QAM &


NG-PON2 RF
Spectrum
Triplexer ATT
: Optical signal Analyzer RF port’s noise level
: Electrical signal after PON path On  Noise level limit <-50dBmV
: Measuring Optical PWR

X-talk Measurement Setup

Page ▪ 125
All Rights Reserved. Copyright©2018 Lightron
New Approach: Test Data & In-house Feedback
✓ Better Crosstalk performance is required
Analog Receiver Specifications
Test Result Remarks
Parameter Min. Typ. Max. UoM
Operating Wavelength 1550 - 1560 nm O.K
Responsivity 0.8 0.9 - A/W 1.0
Operating Frequency 54 - 1002 MHz O.K
RF Output Level 17 - 20 dBmV O.K
RF Frequency Response Flatness - - 2 dB 3.0 Need to improve
RF Output Tilt 0 1.5 3 dB -2.0 Need to improve
RF Return Loss 14 - - dB 12.0 Need to improve
CSO 60 - - dBc 62.2
CTB 60 - - dBc 61.6
CNR 44 - - dBc 40.3 Dependent on X-talk improvement
MER 35 - - dBc 26.6 Dependent on X-talk improvement
Crosstalk (Digital to analog signal) - - 3 dB 17.0 Need to improve

Note 1. Crosstalk improvement remedies Note 2. Values need to be improved


① Analog/digital board separation
② Intensified Ground & EMI Shielding

Page ▪ 126
All Rights Reserved. Copyright©2018 Lightron
Summary & What to do next

Remove gating issues @ Diplexer


- PMD specs : OOC2, OPP
- Class 2 Ch switchover time
- OMCI verification Extend NG-PON2 Offering to
the market
- Class N1 OLT in XFP
Improve RF Performances @ Triplexer - SFP+ ONU with MAC
- X-Talk capability
- CNR
- MER

Page ▪ 127
All Rights Reserved. Copyright©2018 Lightron
Component market update
Access finally has wavelength flexibility
Next Generation Passive Optical Networks 2 (NG-PON2) has challenged the
traditional way of treating access real state, the wavelength, by introducing high-
resolution tunable devices with the tight, 100GHz, channel spacing. This ITU path
brought an extra dimension to the access space allowing flexible management of
the network, opening windows to smooth pay as you grow deployments,
coexistence with all present and close future technologies, as DWDM. This seems
great, however, it pushed the current optical technologies, like GPON, which had
wavelength freedom of tens of nm to tenths of nm (100GHz); from fixed
wavelength to tight control tunable wavelength, and from 2.5Gbit/s to 10Gbit/s x
Antonio Teixeira 4 (or more). To solve this, great technical challenges had to be overcome and even
some compromises had (and have) to be made to foster broad market adoption to
CTO & Founder happen sooner than later. PICadvanced, in this work, will present the current
PicAdvanced status of its solutions and the results of its production units which, given the facts,
that is based on DML and fast thermal control in both Tx and Rx, show great
potential to open the road to broad adoption of the standard.
. 128
Company presentation
• Portuguese startup founded in end of 2014
• Located in Aveiro
• Young and motivated team, growing quickly!

“Think outside the box, with us!”

129
Facilities
University of Aveiro and Instituto de Telecomunicações

130
131

PICadvanced - Who are we?


Main business is focused in optics
Operates through
Development of NGPON2 Components
Innovation
– Specially in Photonic Integration Design and Packaging
Branding and customization

131
132
19.03.18

Our solutions
In-house and OEM products reaching access and core networks!
ONU BOSA
- Class 3 - 1Q 2016
- Class 2 - 1Q 2017

OLT XFP – 2Q 2018

ONU XFP
- Class 2 - 2Q 2017
- Class 3 - 1Q 2017
- Ctemp + Itemp 4Q 2017

132
133
19.03.18

Tunability advent

133
134
19.03.18

Tunability advent – PayG and Soft Net Mngt ...

134
Challenges stemming from tunnability
Amplitude
• STSE – Short Term spectral 1 1 1
excursion
time
– Within Channel
Wavelength
• MSE – Maximum Spectral
excursion
– +/- 20GHz time

• Long term stability


Wavelength
– >5000h
• Fast Fine tuning A lot of time
– Class 2 (<=25ms Tx and Rx) Tunning time

– Class 3 (<=1s Tx and Rx) Ch4

Ch3
Wavelength
• Temperature range stability
Ch2
Ch1

– Commercial Temp (0..70ºC) A lot of time


– Industrial Temp (-40..85ºC) Wavelength

-40ºC 85ºC
135
Operational model
T and I actuators

T Sens.

Filtered Rx T Sens.
driver
T&I routine T&I routine
TEC routine STSE
T Sens. TEC

T Sens.
T and I actuators
mcontroler
TEC routine T&I routine
from device and other
Other parameters

Laser
readings

T Sens.
T Sens.
STSE, MSE, Temp, Long Term and Class 2 operation
Class 3 operation TEC
Driver

136
NG-PON2 ready
Statistics and production robustness ongoing - Transmitter

137
NG-PON2 ready
Statistics and production robustness ongoing - Transmitter

138
NG-PON2 ready
Statistics and production robustness ongoing - Transmitter
CH1 CH2
10 10

0 0

Optical Power [dBm]


-10
Optical Power [dBm]

-10
-20 -20

-30 -30

-40 -40

-50 -50

-60 -60

-701531 1532 1533 1534 1535 1536 1537 -701531 1532 1533 1534 1535 1536 1537

-80 -80
Wavelength [nm] Wavelength [nm]

CH3 CH4
10 10
0 0
Optical Power [dBm]

Optical Power [dBm]


-10 -10
-20 -20
-30 -30
-40 -40
-50 -50
-60 -60
-701531 1532 1533 1534 1535 1536 1537 -701531 1532 1533 1534 1535 1536 1537

-80 -80
Wavelength [nm] Wavelength [nm]

139
NG-PON2 ready
Statistics and production robustness ongoing - Receiver

• Some 1->4 and 4->1 transitions still above 25ms


• W.I.P to be fully compliant.
• FW modifications under implementation

140
PIC based transceivers
Extra functionalities through simplication
Design
• innovative approache brings coherent to PON through optic integration
• Proprietary BB that reduce complexity and floor space on the PIC increasing the potential of low cost integration
• Several iterations of the design already done through MPW runs –mature design

141
PIC based transceivers
Extra functionalities through simplication
Packaging
• costum design approach for our application
• Defined path to reach passive alignment: high scalability potential with simple design PIC packaged in the holder
• Designed to use current optics and electronics production machinery

Costum design High-precision


fabrication

Eletrical packaging

Characterization

142
PIC based transceivers
Extra functionalities through simplication!

Packaging
• Development of termal sensors for
PIC surface temperture control
• Linear sensitivity drift
• Low bulk resistivity
• High termal stability
• Integration with holder

143
PIC based transceivers
Extra functionalities through simplication!

Roadmap with integration will bring:


• OLT and ONU in the same chip
• Higher link budget: N2,E1,E2
• Channel bonding
• Pay as you grow in the optics level
• RoF capabilities
• Multitech transceivers: different technologies in the same PIC

144
Conclusions
• Tunability and all its advantages has arrived to PON

• Challenges of tunability and tight laser control were overcome

• Cost effective solutions for NG-PON2 are available

• PIC solutions and becoming real and will enhance the capabilities of tunability in
the networks

• Market development/growth is required to improve the production and cost


levels of the devices

145
more at picadvanced.com
[email protected]

This work was supported by Fundação para a Ciência e a Tecnologia (FCT) under the project
“COMPRESS - All-optical data compression” – PTDC/EEI-TEL/7163/2014, the QREN/COMPETE
P2020 project “HeatIT” ref. 17942 and the CENTRO2020 P2020 project “Internacionalização” ref.
26018.

14
Component market update
Optical transceivers for NG-PON2 and
beyond: an update
We review the recent progress made in the optical
transceivers for NG-PON2. The key issues to
improve the cost–performance ratio are addressed
David Li and the potential solutions are proposed.
Co-founder and CTO
Hisense Broadband

147
Progress of Optical Transceivers
for NG-PON2 and Beyond

David Li,
Hisense Broadband

March 15, 2018

NG-PON2 Council Workshop, 2018 OFC


Contents
 Introduction
 Progress in NG-PON2 OLT
 Progress in NG-PON2 ONU
 Path to mass production

NG-PON2 Council Workshop, 2018 OFC 149/1


NG-PON2 System
NG-PON2 10G-PON
λd1 … λdn Tunable Tx
λu1 Tunable Rx λ1 Fix Tx
λu1 Fix Rx
TRx1
TRx2 λd1 … λdn 1:N splitterλd1 … λdn Tunable Tx FixTx
Fix Tx
λu2 Tunable Rx λ1
TRx TRx λu1 Fix Rx
Fix Rx
λu1 … λun
3
TRxn 1:4 splitter

λd1 … λdn Tunable Tx 1:N splitter λ1 Fix Tx


WDM 1:N splitter λu1 Fix Rx
λun Tunable Rx

NG-PON2 OLT NG-PON2 ONU 10G OLT 10G ONU

New Features in NG-PON2


 NG-PON2 is a hybrid PON using both TDM and WDM. NG-PON2 is a 40Gbps capable passive
optical access networks and trunk bandwidth is 4 times of 10G-PON.
 NG-PON2 ONUs use the tunable transmitter and receiver.
 In addition to the access applications, 5G front haul is considered as one of the most important
applications of NG-PON2
NG-PON2 Council Workshop, 2018 OFC 2/16
Challenges in NG-PON2
 Technically
 Higher link budget (compensate Mux loss)  Wavelength drift at BM on/off
 Tunable TX and tunable RX in ONU  Interference to the adjacent channel at BM on/off
 OOB noise suppression >45dB  Channel switching time

New devices in NG-PON2


1. 1596~1599nm, high power OLT EML+SOA;
2. 1532~1535nm, high power, low drift and tunable ONU TX;
3. Compact tunable RX
NG-PON2 Council Workshop, 2018 OFC 3/16
NG-PON2 Spec
Wavelength Plan
Channel Wavelength Downstream Wavelength Upstream
ƛ1 1 1596.34nm 1532.68nm
ƛ2 2 159719nm 1533.47nm
ƛ3 3 1598.04nm 1534.25nm
ƛ4 4 1598.89nm 1535.04nm

ONU Tuning Requirements


Class Optic Tuning Time Application
Class 1 <10us Not Used
Class 2 10us to 25ms Enterprise Svcs
Class 3 25ms to 1sec Residential/SoHo
Current residential deployment requires Class 3 only. Yet It is desirable to achieve
Class 2 for enterprise applications. 4/16
NG-PON2 OLT Status
Down Stream
Class G.989.2 Amd1 (TWDM, S/R-CG) (Down Stream) Actual Module Tx
(10G/10G) Path Loss Max OPP max Sens min Power S/R-CP (WM=2.2dB) Power ER (min 8.2dB) Disperion Penalty
N1 29 2dB -28dBm +3dBm +5.2dBm >6dBm >8.5dB
N2 31 2dB -28dBm +5dBm +7.2dBm >7.5dBm >8.5dB <1dB
E1 33 2dB -28dBm +7dBm +9.2dBm >9.5dBm >8.5dB

▪ With existing design of the OLT transmitter, It is possible to support Class N2 and E1 today.

Up Stream Standard & Customer Spec


Class G.989.2 Amd1 (TWDM, S/R-CG) (Upstream) S/R-CP WM Xtalk In-Band System
(10G/10G) Path Loss Max OPP max Sens min Power max WM=2.2dB max 0.5dB Noise (Verizon)
N1 29 1dB -26dBm +4dBm -28.2dBm -28.7dBm -29.2dBm
N2 31 1dB -28dBm +4dBm -30.2dBm -30.7dBm 0.5dB -31.2dBm
E1 33 1.5dB -30.5dBm +4dBm -32.7dBm -33.2dBm -33.7dBm

Upstream Actual Performance


Class OLT Rx ONU Tx (EML+SOA) Additional New Tunable Laser (Target) Additional
(10G/10G) ER=6dB ER>8.2dB Power (4dBm) ER (6dB) Penalty margin Power ER Penalty margin
N1 -29.5dBm >5dBm >8.2dB <1dB
N2 (-31.2dBm) -31.2dBm (+1) (>+1) >2dB
E1 Non >+7dBm >8.2dB 1dB 2dB

▪ If ER=6dB, there is yield loss to reach -31.2dBm. The yield is improved when ER>8.2dB w/ EML+SOA

NG-PON2 Council Workshop, 2018 OFC 5/16


Progress of OLT TRX
Enabling Technologies:
• Transmitter: 10G EML+SOA at 1596-1599nm
• Receiver: 10G high sensitivity APD + burst mode TIA
Status:
• Transmitter meets both N1 and N2 requirements
• Receiver meets N1, but N2 with screening
• Receiver setting time needs to be improved
Specifications Standard Actual Solutions Timeline
Receiver Seek for new APD and TIA to
N2<−31.2dBm N2<−30.5dBm H2,2018
Sensitivity improve sensitivity
Receiver Setting Employ specialized TIA and
<128ns <400ns H2, 2018
Time LIA for XGS-PON

NG-PON2 Council Workshop, 2018 OFC 6/16


OLT Performance

OP(dBm) ER(dB) Margin(%) Wavelength SEN


CH1 7.95 9.51 28 1596.342 -30.5
CH2 7.87 9.27 32 1597.197 -30.3
CH3 7.65 9.68 31 1598.041 -30.8
CH4 7.54 9.6 28 1598.892 -30.6

CH1 CH2 CH3 CH4

OLT output power, extinction ratio, sensitivity, etc. can meet or exceed the N1 standard
requirements – Mass production ready
NG-PON2 Council Workshop, 2018 OFC 7/16
OLT Rx Sensitivity Variation
 TX optical waveform quality

Distorted waveform of DML Good waveform of DML Typical waveform of EML

 Extinction Ratio
With increase of ONU ER, OLT
sensitivity will be improved
• ONU TX waveform quality and higher
extinction ratio will help the
improvement of OLT Rx sensitivity.
• EML is a better choice for ONU TX
• Sensitivity decreased about 2.5dB
when ER changed from 8.2dB to 4dB

8/16
Progress of ONU TRX
Enabling Technologies:
 Transmitter:
1) Thermal tuned DML;
2) Thermal tuned EML+SOA;
3) Thermal tuned EML;
4)Electrical tuned DBR;
 Receiver: 1) Thermal tuned filter; 2) Mechanical tuned filter;
Status:
 Receiver: Thermal tuned filter meets Class 3, MEMS can achieve Class 2;
 Transmitter: +4dBm output power, wavelength shift, TDP
Note: All above options have been evaluated. The best ONU TX approach will be discussed.

NG-PON2 Council Workshop, 2018 OFC 9/16


ONU Laser Technologies for NG-PON2

It is noticed that the DML wavelength distribution per wafer is similar as the EML. The DWDM
packaging of the DML is also the same as EML packaging. The overall cost of DWDM DML package
is not significant lower than that of the EML.

NG-PON2 Council Workshop, 2018 OFC 1580/


Progress of ONU TRX
Test results based on EML+SOA
OP(dBm) ER(dB) Margin(%) Wavelength shift(GHz) TDP(dB)

CH0 7.98 9.01 29 4 <1


CH1 7.20 8.80 28 4 <1
CH2 6.34 8.81 22 10 <1
CH3 6.53 8.88 25 9 <1

CH0 CH1 CH2 CH3

ONU modules can meet higher power and better wavelength shift of by using EML+SOA.

NG-PON2 Council Workshop, 2018 OFC 159/1


Tradeoff between OLT and ONU
 Current OLT BM Rx can support Class N1 w/ margin, but the yield is low to meet the
N2 -31.2dBm spec.
 The OLT sensitivity need to be improved for N2 application in mass production.
 The current commercially available BM TIA is not good enough for NG-PON2. New
NG-PON2 OLT BM TIA is needed to further improving the recovering time.

It is suggested that to meet Class N2, or Class E1 link budget, the best approach is
to apply ONU transmitter with higher ER (>8dB).

NG-PON2 Council Workshop, 2018 OFC 160/1


Improvement of OLT/ONU TRX

Standards Q1/2017 Q1/2018


Min: >4dBm
Optical power Type A: 4-9dBm 0dBm
Typical: 5~8dBm
Burst wavelength shift <20Ghz >20Ghz <15GHz
Operation C: 0~70deg
0~50deg 0~70deg
Temperature I:-40-85deg

 OLT TRX has been delivered about 10Kpcs in 2017


 ONU TRX has been delivered in thousands so far

NG-PON2 Council Workshop, 2018 OFC 13/16


Cost Reduction Trend
As of today, PON have been deployed widely and reached about >70Mpc per year. In the past
ten years, the PON module price dropped to 1/10 with increasing demand and cost reduction. It
is expected that the trend will be the same for10G PON and NG-PON2.

NG-PON2 Council Workshop, 2018 OFC 162/1


Conclusion
 NG-PON2 can provide super bandwidth for fixed access and 5G mobile
applications.
 N1 type of NG-PON2 modules can be delivered in volume.
 The major issues to affect the yield of N2 modules are the receiver
sensitivity and BM recovering time. These issues are being solved with the
improvement of the key components.
 The cost of the NG-PON modules will be reduced further, with the volume
increasing, the improvement of the yield and production efficiency.
A novel tunable, high power, high yield, low cost ONU laser is under
development by the major laser suppliers. It is expected that the new laser
will be available in 2018, which will help the performance increasing and
cost reduction.

NG-PON2 Council Workshop, 2018 OFC 15/16


Thanks
Welcome to visit Hisense show booth #2031
Component market update
Low-Cost Coherent Detection for
NG-PON2
We present the most recent update on our effort to
develop Quasi-Coherent Receiver Technologies that
pave the way towards a realization of the higher NG-
PON2 Classes. THis update will include the first tests
Jesper Bevensee Jensen of the Quasi-Coherent receiver in combination with a
CTO & Founder
commercially avaiable ONU transmitter.
BiFrost Communications

165
Bifrost
COMMUNICATIONS

Low-Cost Coherent Detection for NG-PON2

Jesper Bevensee Jensen, [email protected]

16
Bifrost Simplified Qazi-Coherent Receiver
LO laser
Analogue signal
processing chip
Signal
Combiner
Polarizing Photo
beam splitter detectors

• Only very little equipment (2 PDs, 1 PBS, 1 pcs. 3 dB coupler)


• No DSP (Analog signal processing chip only consumes 10 mW)
• Vast simplification
• Coherent potentially ready for access

[email protected] 167
First tests with Commercial NG-PON2 ONU Transmitter
Channel 0 Channel 1

TX output power
Channel 0 1.47 dBm
Channel 1 0.34 dBm
Channel 2 0.27 dBm
Channel 3 -0.94 dBm
Channel 2 Channel 3

• EML based tunable transmitter


• Very clean TX eye with ER > 11 dB
• SNR varies from 5.49 dB to 8.77 dB
• TX output power highlights NG-PON2
technical challenge regarding ONU power

[email protected] 168
Optical spectrum with and without the LO

LO output power +15 dBm (6.5 dBm at PD)

[email protected] 169
Experimental Validation - testbed

Analog
Signal
processing

• TLS LO for ease and repeatability of tuning (no penalty against DFB with same power)
• 20 and 40 km SSMF transmission
• Polarization aligned 50/50 between the two arms (worst case theoretically)
• PIN photodetectors (same PDs used for comparison with direct detection (DD))

[email protected] 170
Experimental Validation – BER B2B, 20 km and 40 km

• Huge improvement when compared to DD


• 40 km transmission with low penalty
• Penalty higher for channels 0 and 1
• More investigations needed to confirm
and understand
• Difference in BER slope compared to DD is
expected from theory

[email protected] 171
Sensitivity and Penalty Summary
DD Bifrost Receiver
Channel Sensitivity Penalty Penalty
Sensitivity (BTB) Penalty (20Km) Penalty (40Km)
(BTB) (20Km) (40Km)
0 -19.3 dBm 0.3 dB 1 dB -34.6 dBm 1.6 dB 2.6 dB
1 -19.3 dBm 0.1 dB 0.7 dB -34.6 dBm 1.6 dB 2.6 dB
2 -19.3 dBm 0 dB 0.3 dB -34.7 dBm 0.2 dB 0.7 dB
3 -19.3 dBm 0 dB 0.3 dB -34.8 dBm 0.4 dB 0.8 dB

Improvement
Channel
BTB 20Km 40Km
0 15.3 dB 14 dB 13.7 dB • B2B sensitivity better than -34.6 dBm for all channels
1 15.3 dB 13.8 dB 13.4 dB • 0.2 dB to 1.6 dB penalty after 20 km SSMF
2 15.4 dB 15.2 dB 15 dB • Better than 14 dB improvement over DD after 20 km
3 15.5 dB 15.1 dB 15 dB

[email protected] 172
LO detuning tolerance
PRX = -33 dBm PRX = -32 dBm
Channel Channel
0 1 2 3 0 1 2 3
Valid Low
11 GHz 12 GHz 12 GHz 12 GHz 10 GHz 10 GHz 11.5 GHz 12 GHz
IF
Valid High 26.5
24 GHz 29 GHz 29.5 GHz 27 GHz 28 GHz 29.5 GHz 30 GHz
IF GHz
IF 15.5
14 GHz 18 GHz 18.5 GHz 18 GHz 19 GHz 19 GHz 19 GHz
variation GHz
Equivalent 35.5
34 GHz 38 GHz 38.5 GHz 38 GHz 39 GHz 39 GHz 39 GHz
BW GHz

• “Equivalent BW” = tuning range + signal bandwidth


• 40 GHz bandwidth required for burst mode
• We are nearly there, and expect improvement
• Limited by bandwidth of PD

[email protected] 173
Power budget estimations
DD Bifrost Receiver
Power Remaining Power Budget Remaining Power Budget
Channel PTX Power Budget
Budget after SMF after SMF
BTB 20Km 40Km BTB 20Km 40Km
0 1.47 dBm 20.77 dB 15.47 dB 9.77 dB 36.07 dB 29.47 dB 23.47 dB
1 0.34 dBm 19.64 dB 14.54 dB 8.94 dB 34.94 dB 28.34 dB 22.34 dB
2 0.27 dBm 19.57 dB 14.57 dB 9.27 dB 34.97 dB 29.77 dB 24.27 dB
3 -0.94 dBm 18.36 dB 13.36 dB 8.06 dB 33.86 dB 28.46 dB 23.06 dB

• Assumed 0.25 dB/km loss


• Power budget calculated as sensitivity – TX output power
• Remaining power budget already includes measured dispersion penalty
• > 28 dB Remaining Power Budget for all channels (enough for 256 split + CoEx + WDM MUX)

[email protected] 174
Our Next Steps

Integrate our chip


Finalize Prototype
with partner’s
Optics Design Quasi-Coherent ROSA
transceivers

[email protected] 175
Summary
• Coherent Receiver with vastly reduced complexity for access networks
• Huge potential for NG-PON2 – E2 performance within reach
• Paving the way for reduced ONU transmitter cost (VCSELs?)
• 25 Gbps line rates possible with no added receiver cost if NG-PON2
wavelength plan is used

[email protected] 176
Bifrost
COMMUNICATIONS

17
Component market update
Discussion on low cost NGPON2 ONU
solution based on DML

In this topic, we will like to investigate the


possibility to use DML for NGPON2 ONU solution
Ben Chen under G.989.2 Amendment 2, in order to achieve
Technical Director Datacom low cost and to be ready for massive deployment.
& Access Accelink

178
Discussion on Low Cost
NGPON2 ONU Solution
Based on DML
NG-PON2 G.989.2 Amendment 2
• Lower cost DML path to ONU optics compliance: ER vs Power
• Increase allowable ONU Laser turn on time (from 12 ns to 128
ns)
• Allow ONU transmitter to exceed specified power penalty as
long as transmit power is increase accordingly
DML Solution Analysis (1/4)

• Laser Power

Can reach +5dBm in the range 0℃~+70 ℃,

• ER

Can achieve 6dB in the range 0℃~+70 ℃

--- easy to satisfy the standard requirement.


DML Solution Analysis (2/4)

•TOSA structure

Conventional DFB TOSA + Temperature control (TEC,


thermistor, etc.)

•Cost

Compared to EML + SOA solution, DML (DFB) is cheaper.


DML Solution Analysis (3/4)
•Wavelength shift (TX spectral Excursion)

Comparison:
EML + SOA: SOA works as a shutter
DML(DFB): Operating current is around 40-50 mA
EML: LD operating current for LD is around 100mA

There are some hardware/firmware methods to reduce the wavelength


shift.
More Close Look at Wavelength Shift
• DML Wavelength shift (Spectrum extension)
caused by direct modulation mode
• DML wavelength blue shift in burst mode
• Wavelength shift of DML in an ONU may not be
solved easily even if Cooled DML

Is there any ONU


solution with low
cost Uncooled
DML?

Quote from "25G DML Wavelength Shift Measurement for 25G EPON ONU Wavelength Shift Estimate with Burst Mode" by M. Li, H. Zhang, J. Zheng, AOI
DML Solution Analysis (4/4)
•ONU LD Wavelength
Currently not so many commercial 1532nm-1536nm DFB
available → the cost is high.
•Dispersion
Since the ONU LD wavelength is defined at C band, TDP is
high, need high performance DFB → the cost is high

-- The cost is the key for NGPON2 commercial deployment!


Can we consider
other available
DML solution
for NGPON2
ONU?
O-band CWDM DML (1)
L0 L1 L2 L3

CWDM4 MSA Technical Specification

XG(S)-PON GPON

1260 1270 1280 1290 1300 1310 1320 1330 1340 1350 1360
❖ Matured and Commercially available (low cost)
❖ 4 * CWDM DML at O-band (low TDP)
❖ Optional Uncooled DML (low power consumption)
❖ Pay as you grow
Upstream Data rate & channels can be increased on demand, up to 40Gb/s
L0 channel (10Gb/s) is still available even if GPON coexistence is necessary
O-band CWDM DML (2)
Scheme A L0 L1 L2 L3

1271nm TOSA-0

1291nm TOSA-1
O
MUX
1311nm TOSA-2

1331nm TOSA-3

❖ No expensive tuning parts (low cost)


❖ No crosstalk based on CWDM channels (simple)
❖ Scheme A
Simple OSA package (low cost)
O-band CWDM DML (3)
Scheme B
L0 L1 L2 L3
Integrated OSA
1271nm

1291nm O
MUX

1311nm

1331nm

❖ No expensive tuning parts (low cost)


❖ No crosstalk based on CWDM channels (simple)
❖ Scheme B
Highly integrated OSA (compact package)
Comparison
CWDM DML*4ch +
Comparison Tunable_DML DBR_DML Tunable_EML + SOA
MUX
Wavelength O-band C-band C-band C-band
Tuning method No tuning External Cavity Electrical Electrical

TEC (C-Temp) NA Additional Control NA Y

Launch Power in fiber ✔✔✔✔ ✔✔✔ ✔✔✔ ✔✔✔✔✔

Extinction Ratio ✔✔✔ ✔✔ ✔✔✔ ✔✔✔✔✔


TDP ✔✔✔✔✔ ✔ ✔✔ ✔✔✔✔✔

Spectral excursion ✔✔ ✔ ✔ ✔ (NA) ✔ ✔✔✔ ✔✔✔✔✔

Tuning Time ✔✔ ✔ ✔ ✔ (NA) ✔✔ ✔✔✔ ✔✔✔✔

Data Rate Over 10Gbps 10Gbps 10Gbps Over 10Gbps


Summary
• Positive move for G.989.2 Amendment 2 to open the path to
DML solution.

• C-band DML solution still has many issues in term of TDP,


tuning speed, wavelength shift and cost.

• If we consider commercially available O-band CWDM DML, it


may resolve the most issues to achieve the low cost.
Thanks!
www.accelink.com
Lunch Break sponsored by
The workshop will return at 13.10

19
Ecosystem Market update
Segment 3

Moderator: Kenneth Gould


Sen. Director Cable MSO, Calix
19
Segment 3: ECO System overviews and applications (Part 1)

13:10 – 13:30 NG-PON2 Interoperability: Challenges and solution.


Zigi Putnins DMTS, Verizon
This talk identifies the NG-PON2 interoperability challenges that need to be addressed, and discusses the role of Verizon
OpenOMCI and Open NETCONF/YANG specifications to solve these challenges. It follows with an overview of the Verizon NG-
PON2 interoperability verification program that has been running for nearly a year and a half in Waltham, MA, and discusses the
results achieved so far along with the plans for 2018.

13:30 – 13:50 NG-PON2 for radio access network


Jun Terada Senior Research Engineer, NTT
NG-PON2 supporting radio access network of 5G/5G+ is presented as a new market of NG-PON2. Requirements for mobile
network and a new interface for low latency transmission are explained and a demonstration is introduced.

13:50 – 14:10 NGPON2, Convergence and OSP Architecture


Erik Gronvall VP of Service Provider Strategy and Networks, CommScope
NGPON2 brings great advantages and opportunities to service providers. With this consideration need to be made for OSP
design with introduction of NGPON2 and wireless/wireline/IOT/business services convergence. The importance of flexibility
throughout the OSP and requirements for power for edge devices adds complexity to the engineering of networks.

14:10 – 14:30 “Evolution of FTTx Networks in a Mobile World”


Kevin Bourg Director, Optical Network Architect, Corning
The convergence of both wireless and wireline technologies is becoming widely accepted as the future-ready platform to meet all
subscriber needs. With NG-PON2 technologies evolving the bandwidth capabilities of fiber networks, ensuring the right
infrastructure is in place in the passive plant is paramount. The level of densification 5G networks demand will necessitate a fiber
network so prolific

195
Segment 3: ECO System overviews and applications (Part 2)

14:30 – 14:50 Driving access network convergence through next-generation PON and SD-Access
Ryan McCowan Director Product Management, Adtran
As operators push more and more fiber deeper into the network to support a range of residential, enterprise and x-haul
services, the need for access network convergence is greater than ever. By combining leading access technologies like
NGPON2 with SDN-based programmability, operators are now able to build highly programmable and scalable converged
access networks. This presentation will cover the use cases driving convergence and the best practices for implementing
these converged, SDN-controlled access networks.

14:50 – 15:10 Challenges and Solutions for an NG-PON2-based Fronthaul


Luca Valcarenghi Scuola Superiore Sant’Anna, Pisa, Italy
The New Radio Access Architecture (New RAN), as defined by 3GPP in TR 38.801, features the disaggregation of the next
generation node B (gNB) into a Distributed Unit (DU) and Central Unit (CU). The distribution of functionalities between CU
and DU impacts the characteristic, in terms of capacity and latency, that the CU-DU connection (i.e., the fronthaul) shall
guarantee. This presentation shades some lights on the suitability and on the attentions to be taken when an NG-PON2 is
utilized to carry fronthaul traffic.

15:10 – 15:30 The Effect of Virtualization on PON Architectures and Hardware


Ed Boyd, CTO and Cofounder, Tibit Communications
Today’s PON systems are massively integrated and application specific systems that, in addition to the PON functions,
contain multiple layers of switching and traffic management. PON systems are becoming virtualized and the impact on
system hardware is subtle, yet significant. In this presentation we will compare the integrated and virtualized architectures
and explore the advantages of virtualized PON.

196
ECO System market update
NG-PON2 Interoperability: Challenges
and solution.

This talk identifies the NG-PON2 interoperability challenges


that need to be addressed, and discusses the role of Verizon
OpenOMCI and Open NETCONF/YANG specifications to solve
Zigi Putnins these challenges. It follows with an overview of the Verizon
Distinguished Member of NG-PON2 interoperability verification program that has been
Technical Staff, Verizon running for nearly a year and a half in Waltham, MA, and
discusses the results achieved so far along with the plans for
2018.
197
OFC NGPON2 Council Workshop
NG-PON2 Interoperability:
Challenges and solution
Z. Putnins
DMTS, Verizon
3/2018

Confidential and proprietary materials for authorized Verizon personnel and outside agencies only. Use, disclosure or distribution of this material is not
permitted to any unauthorized persons or third parties except by written agreement.
Agenda

1. Lessons learned with BPON and GPON


2. NG-PON2 Interop
3. Verizon Open OMCI
4. Verizon Open NETCONF/YANG
5. Interoperability Validation

199
Lessons Learned
NG-PON2 – third generation of PON systems in Verizon
▪ BPON deployment commenced in 2005: Tellabs, Motorola
▪ GPON deployment commenced in 2008: Alcatel-Lucent, Motorola
▪ NGPON2 deployment commencing in 2018: Calix, ADTRAN

Does after-the-fact Interop work?


▪ Vendor builds business case assuming deploying both OLT and ONT (closed system)
▪ No inherent motivation for standards compliance
▪ Even if a vendor claims standards compliance, many standards support
▪ Optional items
▪ Vendor-proprietary extensions
which leads to standards-sanctioned non-interoperability
▪ Even if a vendor develops, in good faith, a standards-compliant product, there is not real way to verify compliance as part of
black-box (closed system) ONT/OLT testing

200
Going forward: NGPON2

NGPON2 RFP issued at the end of 2015


▪ Interoperability requirements formulated upfront and are applicable at day one of deployment
▪ Cost models for complete system, as well as individual components
▪ Vendors agree to compliance with Verizon interoperability specifications, and participation in Verizon
sponsored interoperability test program
▪ NG-PON2 interoperability has to be there from day one

Verizon established “Verizon Open OMCI” group, consists of system vendors and SoC
vendors, to define specifications.
▪ Typically meets every week to discuss interop process/issues, documentation review etc.

201
Going forward: NGPON2 (2)
Verizon OSS
Interoperability is required between:
▪ ONT and OLT CT/EMS from different vendors within a wavelength channel EMS
▪ Instances of OLT CT/EMS from different vendors within an NG-PON2 system
▪ OLT CT and ONT with optical transceivers of different vendors ONT A
OLT CT C
▪ OLT CT/EMS and ONT from different vendors with the functional elements of
Verizon network architecture

ONT C

G.988
2017/2018 OMCI OMCI
G.989.3
TC TC
2016 and early 2017 G.989.2
PMD PMD

202
Going forward: NGPON2 (3)
Three pillars of NGPON2 Interop
1. Specifications: includes application use cases, sequence flows, ME definition (if needed), attribute definition &
disambiguation (if needed).
▪ Updates to the specification are published once agreement is reached within the Verizon Open OMCI group
2. Validation: using a combination of vendor self-certification and Verizon-observed cross-vendor interoperability
demonstration to validate specification compliance and correct functional operation.
3. Deployment: go through the formal Verizon product verification process and appropriate OSS changes needed to support
interoperable deployments as BAU
▪ Currently a work in progress

203
What are we trying to interoperate?
Functions Goal #1: swap out one Vz specification
compliant ONT with another with no
▪ Service definition and performance degradation in performance, features, function
▪ Alarms and PMs or operational support.

▪ Software download (SW image depends on ONT)


Goal #2 any interop issue discovered in the field
▪ Operator and vendor maintenance and diagnostics has same level of impact as a “sw bug”.

Examples of changes/clarifications to G.988


▪ Makes necessary extensions to support multi-wavelength architecture of NG-PON2 and new features introduced by
G.989.2/.3 (new TWDM MEs)
▪ Move to all 64-bit PMs, where defined
▪ Specify ONU2-G Equipment ID is the product’s CLEI code
▪ Disallow any “default value” for capabilities; disallow use of any deprecated MEs or attributes
▪ Restrict SIP configuration to “file retrieval” for VoIP Services
▪ Enhance remote debug ME to accept commands more than 25 characters
▪ Require support for 2 instances of Ethernet History PMD (binned and running counts)

204
Specification: Verizon OpenOMCI
A specification being developed by Verizon
▪ Is based on G.988, as amended, with best practice Appendices. Includes both a narrative word document and a spreadsheet
precisely identifying required MEs, attributes, attribute ranges, alarms, etc
▪ Defines end-user applications, and for each application, defines the supported OMCI MEs
▪ Service profiles (similar to Bluetooth profiles) – what MEs ave to be supported for a specific application (SIP voice, Internet
access, etc.)
▪ No optional MEs; No optional attributes
▪ No vendor-proprietary OMCI objects*.
* Proposed extensions are discussed in the Verizon OpenOMCI committee, and if accepted, adopted as “vendor specific” MEs.
Such MEs are documented with semantics, methods, relationship diagrams, and message sequences, and are designated as
“vendor specific” until adopted by ITU (ex: OMCI support for TWDM)

Version 1.0 has been released, presented to Q2 which accepted the core MEs into G.988 (a subset of
Verizon OpenOMCI).
And published on the Verizon website
http://www.verizon.com/about/techspecs/verizon-openomci-specification

205
Specification: Verizon Open NETCONF/YANG
Another specification being developed by Verizon
▪ To leverage existing deployments of Layer 2 Network Interface Devices (NIDs), ONTs supporting advanced Layer 2 services
are dual-managed
▪ The advanced Layer2 services are managed using NETCONF protocol and YANG models that cover NID traffic management,
Service OAM, Service Activation, and other Verizon-specific operational features.
▪ NETCONF protocol and YANG model definition went through same process as the Verizon Open OMCI through the Verizon O

206
Validation: Verizon NG-PON2 Interoperability Testing
Program
Based in Waltham, Mass
▪ ADTRAN, Broadcom, Calix/Ericsson, Cortina Access

Interop Events conducted bimonthly


▪ 2016 – two events ; 2017 – six events; 2018 – one event completed, 5
more planned
▪ Covers optical, PMD, PLOAM, OMCI interoperability for basic PON
functions and for Single Family Unit services. Covers NETCONF/YANG for
NID services.
▪ Since August: includes the operational aspect of interoperability and
involves Verizon Operations and Engineering as an observers for the plug
and play test scenarios

Vendors travel on-site for 4 days of interop testing, following a Verizon-authored Test plan
Vendors submit a test result document; any discrepancies are logged and appropriate corrective
actions tracked

207
Dimensions of NG-PON2 interoperability

Cross vendor ONT on


same lambda
Different
vendors OLT on
same PON

Cross OLT-vendor
TWDM mobility with
consistent ranging and
no MIB reload.

208
Accomplishments
1. Cross vendor ONT Activation (cross vendor: OLT and ONT from different vendors)
2. Cross vendor service activation with same level performance across ONTs
3. Cross vendor SW download
4. Cross vendor PM collection and alarm generation
5. Cross vendor TWDM mobility with consistent ranging and no MIB reload
6. Replace running ONT with other-vendor factory fresh ONT and restore service
7. Same DBA performance for both uncongested and heavily congested PONs
8. Cross vendor Layer 2 service activation using NETCONF/Yang
9. Identified work-arounds for SoC issues, raises the issue that interoperability extends to
addressing defects
10. Identified ambiguities in standards, and forward to appropriate body

209
Verizon 2018 NG-PON2 Interoperability Sessions

Interop event Scope


Feb 2018 OpenOMCI compliance and interoperability for LCI and IBONT.
Apr 2018 OpenOMCI compliance and interoperability for LCI and IBONT. Open NETCONF/YANG
for IBONT provisioning. IPFIX data collection

Jun 2018 OpenOMCI compliance and interoperability for IBONT. Open NETCONF/YANG for
IBONT provisioning. IPFIX data collection

Aug 2018 Open NETCONF/YANG for IBONT provisioning. IPFIX data collection
Ethernet OAM for LCI service (tentative)
Oct 2018 ICTP
Dec 2018 TBD

LCI – internet-only service


IBONT – ONT with integrated Advanced Layer Services

210
Conclusion

Despite appearances, we are not trying to


define a new standard
▪ Use existing standards.
▪ Disambiguate to meet Verizon’s service deployments
▪ Extend standards and submit addendums to relevant
bodies

Get vendors in the lab, and verify interoperability at the protocol (syntax)
and functional (semantic) level, for both service and operational support

211
Thank you.

Confidential and proprietary materials for authorized Verizon personnel and outside agencies only. Use, disclosure or distribution of this material is not
permitted to any unauthorized persons or third parties except by written agreement.
ECO System market update
NG-PON2, for Radio Access Network

NG-PON2 supporting radio access network of 5G/5G+ is


presented as a new market of NG-PON2. Requirements for
mobile network and a new interface for low latency
Jun Terada transmission are explained and a demonstration is
Group Leader - Optical Access introduced.
Systems Design Group, NTT
Access Network Service
Systems Laboratories
213
NG-PON2 for Radio Access Network

15 Mar., 2018
NTT Access Network Service Systems Labs.
Jun Terada

Copyright©2018 NTT corp. All Rights Reserved.


Outline

1. Trends in Mobile Services

2. 5G Mobile Network and Mobile Optical Network

3. RAN Evolution and Future MFH Network

4. NG-PON2 for 5G MFH Network

5. Summary

Copyright©2018 NTT corp. All Rights Reserved. 215


Trends in Mobile Services
• Mobile throughput growing rapidly. By 2020, 1000x that of 2010.
• NTT docomo started 788 Mbps service in September 2017.
• 5G service is expected to start in 2020 and some R&D projects have started to realize
5G mobile networks.
3.9G 4G(LTE-
2G 3G 3.5G (LTE) Advanced)
5G
10G Now
Throughput (bit/s)

5G
LTE-Advanced
1G LTE > 10 Gbit/s
LTE 150 Mbit/s375 Mbit/s
LTE 100 Mbit/s 4G
100M 75 Mbit/s
HSDPA HSDPA 3.9G
7.2
3.6 Mbit/s Mbit/s HSPA
10M
14 Mbit/s
WCDMA 3.5G
1M 384 kbit/s
3G
2000 2010 2020
Year
Copyright©2018 NTT corp. All Rights Reserved. 216
Architecture of Base Station
• Currently, there are 2 base-station architecture:
Distributed RAN and Centralized RAN.
Core network Core network

Mobile backhaul (MBH)


Mobile backhaul (MBH)

BBU BBU BBU

Mobile fronthaul (MFH) CPRI


BBU BBU BBU
RRH RRH RRH RRH RRH RRH

Distributed RAN Centralized RAN


• MFH requires 16 times larger bandwidth than wireless throughput.
• Current MFH of Centralized RAN is P2P connection.
Copyright©2018 NTT corp. All Rights Reserved. 217
Functional split between CU, DU and RU
High Low High Low High Low
PDCP RF
RLC RLC MAC MAC PHY PHY
layer TRx
layer layer layer layer layer layer
Higher Layer Split (HLS) Lower Layer Split (LLS) CPRI
F1 Interface eCPRI
CU DU/RU
CU DU RU
CU/DU RU

CU/DU
Low High Different split for different use case
PHY PHY PDCP

Rural Dense urban V2X


Wide coverage RU High CoMP RU Low DU/RU Low latency
performance PHY RLC
RF
RF

Copyright©2018 NTT corp. All Rights Reserved.


218
5G RAN and Mobile Optical Network
• Mobile Optical Network should accommodate various types of RATs
that have various requirements to the network.
New Radio
(Spot cell)

Small cell

Macro cell

Central Optical fiber


office etc.

Mobile Optical Network


Copyright©2018 NTT corp. All Rights Reserved. 219
Resource Block in Mobile System

• Each UE sends upstream data (PUSCH data) in 1 time slot (TTI)


according to resource block allocated by base station.
• Base station allocates individual resource block element to each UE
so as not to be allocated one element to multiple UEs.
Base Station

PDCCH
PUSCH
UE1
UE2
UE3

PDCCH
- Downlink control information
- Downlink grants
1 TTI 1 TTI 1 TTI - Uplink grants
Resource block element
Resource block allocated to each UE - Uplink power
Copyright©2018 control
NTT corp. All Rights Reserved. 220
Fronthaul for Various types of Base Stations
• Current base station has an individual resource block and fronthaul
link and multiplexing some links requires large bandwidth as well as
Massive MIMO antenna system with high frequency.
• Distributed antenna system uses a single resource block shared by
multiple RUs and suitable for TDM-PON such as NG-PON2.

DU DU DU DU DU

RU RU RU RU RU RU RU
Large resource block Single resource block shared
Individual resource block for high frequency with distributed antennas
for each base station with massive antennas Copyright©2018 NTT corp. All Rights Reserved. 221
Utilization of NG-PON2 in RAN
• NG-PON2 can be utilized for a base stations with any splitting points.
• Location of CU/DU/RU for Distributed Antenna System
– CU at central office
– DU and RU at High-density area (stadium, office, shopping mall, factory etc.)
• F1 interface may not require low-latency transmission.

Stadium Office
NG-PON2 RRH
NG-PON2!
with low latency
CU DU RU
Higher Layer Split Lower Layer Split
(HLS) (LLS)
Copyright©2018 NTT corp. All Rights Reserved. 222
Low-latency NG-PON2 for MFH network
• Statistical multiplexing of packetized MFH achieves effective use of
optical bandwidth.
• Big issues to use PON technologies is latency.
– Large uplink latency because of conventional SR-DBA Latency
caused by TDM access
SR-DBA based scheduling
DU
RU PON Wireless
Data

OSU(OLT)
ONU Scheduling
ONU
Splitter OLT
ONU Request Grant

DU/
ONU Wireless
ONU Shared Fiber High
ONU Data
RU Latency
100~200 m 5~10 km Wireless
Signal Time
UE Copyright©2018 NTT corp. All Rights Reserved. 223
Cooperative IF with DU
• Cooperation between mobile and optical scheduling reduces uplink
latency.
– Wireless scheduling based PON scheduling.
• The cooperative IF between DU and OLT is required and
standardization is necessary for wide use.
DU Wireless
Scheduling
Cooperative IF DU
OLT PON Wireless
Scheduling Data
Grant OLT
Grant
ONU ONU ONU ONU Wireless Low
RU RU RU Data Latency
RU Wireless
Signal Time
UE UE Copyright©2018 NTT corp. All Rights Reserved.
T. Tashiro et al., OFC2014
224
Functions for Cooperation
• Mobile system and optical system use different languages.
• OLT needs an interpreter that translates mobile language into PON language to
generate bandwidth map (Time and size for each ONU uplink)
• This kind of cooperative DBA is currently being discussed in ITU-T SG15.
• Interface for BBU side might be standardized for various types of BBU with
different splits in near future.

BBU Mobile system RRH


Scheduling
Main signal information for
each UE
Interpreter Main signal
Grant
size
BWmap
generator
OLT IF ONU
TDM PON Copyright©2018 NTT corp. All Rights Reserved. 225
Effect of Cooperation
• Latency without Cooperation is around 1 ms.
• Cooperative DBA through Cooperative IF Lowers the latency to
under 50 µs.

1400
1300 w/o Cooperation with BBU
1200 1 ONU
1100 3 ONUs
4 ONUs
Latency [ms]

1000
900
800
700
600
500
50 w/ Cooperation with BBU
400
40
300
30 1 ONU
2 ONUs
200
20 4 ONUs
100
10
00
0 1000 2000 3000 4000 5000 6000
Copyright©2018 NTT corp. All Rights Reserved. 226
Data rate (each ONU) [Mbps]
E2E Demonstration with Low-latency NG-PON2
• E2E demonstration from App. Server to User Equipment using low-latency NG-
PON2 was demonstrated at NTT R&D forum 2018 in mid Feb.
• HLS/LLS Base Stations coexist in one PON branch.
http://www.ntt.co.jp/news2018/1802e/180214a.html
Server PC
(PC 0) Video streaming
Coop. Coop.
UE &
function function LLS RU
ONU App PC
LLS CU/DU (PC2) Coax
(PC3)
(PC1)
OLT Video streaming

HLS CU
Other UE
& Server ONU HLS DU/RU
Coax & App PC
(w/o Coop.)
Copyright©2018 NTT corp. All Rights Reserved. 227
Issues for sustainable usage of NG-PON2 in RAN

1. Evolution of Mobile System is very rapid. Transmission rate of LLS


such as eCPRI will increase to 25G soon.
PON should support it and migration (from 10G to 25G/50G) should
be easy and cost effective. It is better that replace for the migration is
a small part or module such as optical transceiver, not a large slot of
the system.

2. As for coexisting with other systems such as IoT and FTTH, gap of
transmission rate and traffic volume comparing with such systems
are very large.
This must be considered. One solution is coexistence of multiple-rate
systems (ex. 10G for l1-3 and 25G for l4 in a single system).
Copyright©2018 NTT corp. All Rights Reserved. 228
Summary

1. 5G Mobile network and Mobile Optical Network

2. Functional Split of Base Station

3. NG-PON2-based MFH Network


• Coexistence of HLS/LLS Base stations.

4. Issues of NG-PON2 for continuous use

Copyright©2018 NTT corp. All Rights Reserved. 229


Thank you for your attention!

Copyright©2018 NTT corp. All Rights Reserved. 230


ECO System market update
NG-PON2, Convergence and OSP
Architecture

“NGPON2 brings great advantages and opportunities to


service providers. With this consideration need to be made
for OSP design with introduction of NGPON2 and
wireless/wireline/IOT/business services convergence. The
Erik Gronvall importance of flexibility throughout the OSP and
VP Strategy and Market requirements for power for edge devices adds complexity to
Development, CommScope the engineering of networks “

231
NGPON2, Convergence and OSP
Architecture
How to make networks work together

Erik Gronvall
VP Strategy and Business
Development
Fiber connectivity
for mobile, residential, and
enterprise services

Unified
network Integrated
infrastructure usage of
and Licensed &
architectures unlicensed
spectrum
to improve
efficiency
C-RAN: Fronthaul,
mid-haul, and
WIRELINE backhaul WIRELINE
NETWORK
CONVERGENCE
233 PRIVATE AND CONFIDENTIAL © 2018 CommScope, Inc
Inc.
LAA
A Journey of Transformation
Convergence: Merging of Wireless and Wireline

Drivers
Cell Densification New Spectrum Enabling Technologies New Services

Roadmap

Results
SCALABILITY SPEED SAVINGS

234 PRIVATE AND CONFIDENTIAL © 2018 CommScope, Inc


Inc.
Convergence: The Merging of Wireless & Wireline
Start from shared assets and work towards fully integrated systems

3
Multi-Access Edge Computing
Converged •

Network functions converge
Common credentials & policies
Applications NFV CORD • Containerization
Network Slicing

2
Converged NGPON2 SDN • Common hardware for SDN/NFV
Networking • Applications run on COTS
C-RAN v-OLT v-BBU • Similar processes
Platforms COTS

1
Densification of the wireless network & fixed
Converged •
broadband deployments
Access • Leverage existing footprint and locations for
multi-use, multi-service delivery
Infrastructure fiber cabinets/hubs closures locations • Plan and build wireless/wireline simultaneously

Efficiency, agility, scale, and time-to-market all improve when networks and systems are converged.

235 PRIVATE AND CONFIDENTIAL © 2018 CommScope, Inc


Inc.
Converged Access Network and what it means for NGPON2

• Networks should to be designed and constructed keeping


cost, flexibility and capacity in mind Businesses
• Media installation is a cost driver, resulting in a parallelization
to serialization conversion
splitter

Multiple Fibers

1:4

• Multiple Wavelengths
• Converged Residential/Business/Mobile Services FTTH
• Higher Data Rates

xWDM

Optical
G/EPON Distribution Frame
OLT (ODF)
Multi-Use
Flexible Access Small Cells
NGPON2 Distribution Terminal
Point xWDM

5G C/DWDM

FTddP
CU/DU

5G Fixed Wireless

Central Office/C-RAN hub Distribution Access Sites


236 PRIVATE AND CONFIDENTIAL © 2018 CommScope, Inc
Inc.
LAA
NGPON2 and Convergence in the CO/C-RAN hub
CO and CRAN Locations
• Coexistance elements
Businesses
• Density of fibers
• More Locations (CO’s, Towers, MSC, Cabinets) splitter

1:4
• Handoffs to Wireless
• Edge Compute FTTH
MPO
ODF

xWDM

Optical
G/EPON Distribution Frame
OLT (ODF)
EHD
Multi-Use
Flexible Access Small Cells
NGPON2 Distribution Terminal
Point xWDM

5G C/DWDM tap/monitor

FTddP
CU/DU

5G Fixed Wireless

Central Office/C-RAN hub Distribution DWDM Access Sites


237 PRIVATE AND CONFIDENTIAL © 2018 CommScope, Inc
Inc.
Flexibility for NGPON2 and other services

Flexible Distribution Points


Businesses
• Add/Drop services
• Reconfigure OSP to reach splitter

Businesses / Residences and Cells

1:4
• Allow for crosshaul
FTTH

xWDM

Optical
G/EPON Distribution Frame
OLT (ODF)
Multi-Use
Flexible Small Cells
LTE Access
BBU
Distribution Terminal
Point xWDM

5G C/DWDM

FTddP
CU/DU

5G Fixed Wireless

Central Office/C-RAN hub Distribution Access Sites


238 PRIVATE AND CONFIDENTIAL © 2018 CommScope, Inc
Inc.
Converged Networks will provide leverage

Flexibility is Key
• Fiber Counts Businesses

• Test Access splitter

• Optical Components

1:4
• Locations for Installation
FTTH

Optical xWDM
G/EPON Distribution Frame
OLT (ODF)

LTE
Multi-Use
Flexible Access Small Cells
BBU
Distribution Terminal
C/DWDM
Point
5G xWDM
CU/DU

FTddP
5G Fixed Wireless

Central Office/C-RAN hub Distribution Access Sites


239 PRIVATE AND CONFIDENTIAL © 2018 CommScope, Inc
Inc.
Key Network requirements for NGPON2 and Convergence

Businesses

splitter

1:4
FTTH

xWDM

Optical
G/EPON Distribution Frame
OLT (ODF)
Multi-Use
Flexible Access Small Cells
NGPON2 Distribution Terminal
Point xWDM

5G C/DWDM

FTddP
CU/DU

5G Fixed Wireless

Central Office/C-RAN hub Distribution Access Sites


240 PRIVATE AND CONFIDENTIAL © 2018 CommScope, Inc
Inc.
LAA
Thanks
Reducing latency: Where to place C-RAN (BBU/MEC) hubs?

BBU
BBU
BBU
Backhaul (IP) MEC Fronthaul (CPRI/eCPRI/xRAN)

MSC/Central Office BBU/MEC Cell site


hub

Closer to the core Closer to the edge


• Consolidates a large number • Supports a smaller number of sectors
of sectors
• Leverages existing sites – huts, shelters, cabinets
• Resembles a small data center:
power, cooling, cabling, COTS • Increased requirements for power and fiber

• Incorporates SDN, NFV,


virtualization

242 PRIVATE AND CONFIDENTIAL © 2018 CommScope, Inc


Inc.
ECO System market update
NG-PON2, Enabling new architectures
with converged technologies

The convergence of both wireless and wireline technologies is


becoming widely accepted as the future-ready platform to
meet all subscriber needs. With NG-PON2 technologies
evolving the bandwidth capabilities of fiber networks,
Kevin Bourg ensuring the right infrastructure is in place in the passive
Director, Optical Network plant is paramount. The level of densification 5G networks
Architect, demand will necessitate a fiber network so prolific
Corning

243
Enabling new architectures with
converged technologies
Convergence| Coexistence Element | CapEx Avoidance

Kevin Bourg
Director – Optical Network Architect
Corning Optical Communications

March 22nd, 2018


Agenda
• Convergence of connectivity and speed
• Network Migration: GPON to NG-PON2
• Inside Plant Migration: GPON to NG-PON2
• CapEx Avoidance: An analysis of convergence

Optical Communications © 2018 Corning Incorporated. 245


Convergence of Two Laws
Metcalfe’s Law Nielsen’s Law
V ~ N2 Billboard speeds grows 50% per year
V = value
N = number of connected
Network
devices

Connectivity
1980’s era disk drive
Speed
Convergence

Cisco VNI1 estimates by 2020 there will be 26.3 billion


network devices with wireless accounting for 66% of
devices [1] “The Zettabyte Era: Trends and Analysis, July 2016”, Cisco Visual Network Index
Optical Communications © 2018 Corning Incorporated. 246
Network Migration: GPON to NG-PON2
GPON
ONT

GPON

1:2 1:32

1:2

RF Video GPON
ONT

1:32

OTDR

Optical Communications © 2018 Corning Incorporated. 247


Network Migration: GPON to NG-PON2
MUX
GPON
ONT

NG-PON2

XGS-PON
ONT

GPON
NG-PON2
1:2 1:32 ONT
CEX 1:2
XGS-PON
NG-PON2
ONT

RF Video GPON
ONT

1:32
XGS-PON
OTDR ONT

MUX

PTP
PTP ONT

Optical Communications © 2018 Corning Incorporated. 248


Inside Plant: GPON
Actives Passives

Legend
2X2
Splitter

Stubbed
Housing

Transition
OLT Actives
Network Splice
Equipment
Frame(s)

GPON

OSP
Optical Communications © 2018 Corning Incorporated. 249
Inside Plant: GPON to XGS-PON Migration
Actives Passives

Legend
2X2
Splitter

Stubbed
Housing
XGS-PON
OLT

CEX
Transition Housing
OLT Actives
Network Splice
Equipment
Frame(s)

GPON

OSP
Optical Communications © 2018 Corning Incorporated. 250
Inside Plant: XGS-PON to NG-PON2 Migration
Actives Passives

Legend
2X2
NG-PON2
Splitter
OLT

Stubbed
Housing
XGS-PON
OLT

CEX
Transition Housing
OLT Actives
Network Splice
Equipment
Frame(s)
WM1
GPON
Housing

OSP
Optical Communications © 2018 Corning Incorporated. 251
WDM/CEX form factors

Modules & Housings Cassettes & Housings Splice Trays & Shelves

Optical Communications © 2018 Corning Incorporated. 252


Network convergence: NG-PON2 Scenario
Assumptions
• 100,000 home and mixed use area
• FTTH Build ~$900 HP, ~$600 HC
• 70% aerial, 30% underground
• 1x64 total split centralized or distributed

Potential additional services


• Wireless
• Mobile: Macro cell, Small cell
• Fixed: LoS, Non-LoS
• DAS: In building, Outdoor, Wi-Fi
• Business services (Pt-Pt)
• Back Haul, Interoffice

Optical Communications © 2018 Corning Incorporated. 253


Network convergence: Payoff!
• Although difficult, converging multiple applications onto a single network enables
greater economies of scale
• Example: one system for schools, government offices, traffic lights, security
cameras, Wi-Fi in the city center and residential high-speed internet
Cumulative Carrier CAPEX
$110
Additional Services by DA
$100 • Small cell 25%  $10M
$90
• SME Services 20%  $1M
FTTH • Backhaul 10%  $0.7M
$M

$80 CAPEX for a typical build for 100k home


area with a fully enabled multi-service
$70 network • Cost avoidance of parallel networks results in
$60
15%-50% CapEx savings and significant speed to
Year 1 Year 2 Year 3 Year 4 Year 5 Year 6 revenue advantages
FTTH Small cells Business Services Backhaul Source: Corning Analysis

Converged networks allow addressing multiple market segment simultaneously, adding


revenue streams and de-risking the business case for optical fiber
Optical Communications © 2018 Corning Incorporated. 254
Summary
• Referencing Metcalf’s and Nielsen’s law, we have reached a
point where operators must consider not only the speed
provided to a subscriber but also ensuring ubiquitous
connectivity.
• The migration of a PON network from today’s technologies to
those allowing for a fully converged architecture may be
achieved with little impact to the headend.
• The payoff an operator will experience is significant long-
term capital avoidance.

Optical Communications © 2018 Corning Incorporated. 255


Optical Communications
http://www.corning.com/opcomm

@CorningOpComm
ECO System market update
Driving access network convergence
through next-generation PON and SD-
Access
As operators push more and more fiber deeper into the
network to support a range of residential, enterprise and x-
haul services, the need for access network convergence is
greater than ever. By combining leading access technologies
like NGPON2 with SDN-based programmability, operators are
Ryan McCowan now able to build highly programmable and scalable
Director Product Management, converged access networks. This presentation will cover the
Adtran use cases driving convergence and the best practices for
implementing these converged, SDN-controlled access
networks.
257
Drivers for Disruption

• Need for user-driven, on-demand service delivery


• Improve time-to-market for new products and
services
• Increase network scale and agility to adapt to rapid
changes
• Build intelligent networks with lower operational
costs
• Create a network of interchangeable parts to avoid
vendor lock-in and facilitate innovation

General Business
258
258
Evolution of Broadband Networks
ATM Packet Networks
• Initial launch of
broadband networks
• ADSL2+ DSLAMs and Transition to IP/Ethernet Networks
early MSANs
• Built to scale and drive
down cost Open and Programmable SD-Access
• Driven by transition to
VDSL2, PON and Networks
improvements on core IP • Application of SDN and NFV to enable end-to-end
routing programmability of software centric networks
• Initial focus on next gen access: 10G PON, FTTdp,
DCA, 5G

1990s 2000s 2016 and beyond

General Business
259
Disruption in Network Architecture

Customized, vendor-specific New Paradigm


solutions dominated by 2 or 3
Performance / Adoption

vendors in each segment Open, software-centric


networks optimized for
multi-vendor integration

Period of Discontinuity

Current Regime

Time
Today
General Business
260
Applying Data Center Architectures

• Utilize SDN to create highly


automated and programmable
networks
• Apply NFV to create agile,
software-centric networks
• Disaggregate monolithic systems
• Implemement open networking
architectures

General Business
261
Central Office as a Data Center
Cloud Management System Open, app-
based SDN
SDN Orchestration/Control & NFV MANO control and VNF
orchestration
solution
Scalable data Leaf-Spine
center fabric Switches
network
MetroE Open and
OLTs
PON

programmable
Open and
optical transport
programmable ROADM
access devices
Access I/O Compute and Storage
vBNG
Virtualized
vOLT vRG functions on
commodity data
FTTN DSLAMs, G.fast Virtualized center servers
DPUs, etc. General Business Network Functions
262
Current Network Architecture

Service Provider Orchestration Systems (complex)

Unique APIs per device


High number of
Unique management platforms for each vendor unique integration
points
Transport VendorBB
CPE Vendor Transport
Access Vendor
Vendor B
B Transport Vendor B
Router Vendor B Transport
TransportVendor
VendorBB

Central Office
xDSL
Access Node FTTN Access
Access
Vendor-specific Access
Node Edge Optical
Node
Nodes Router Transport
chassis system Vendor-
Vendor-
Vendor- Vendor- Vendor- Metro
FTTP specific
specific Transport
specific
chassis specific specific
chassis
chassis chassis chassis
system
system
system system system
Fixed Fixed Wireless Access
wireless
Software closely coupled with
General Business vendor-specific hardware
263
Network of the Future

Service Provider Orchestration Systems (simplified)

Open, standardized APIs

Multi-vendor, multi-technology SDN controller

Open, standardized APIs


Central Office / Hub Site
xDSL Open, modular
FTTN Open, modular
Open, modular
access node access
Open, node
modular
Open,node
access modular
access node
access node Modular fabric
Open, modular
switches
Modular fabric
Open, modular Metro
switches
fabric switches transponder Transport
VNF VNF
Open, modular
FTTP Compute and
access node
storage

Software decoupled from hardware


General Business
264
NGPON2 System Architecture
Global Control & Orchestration
Customer Mgmt Control Orchestration
Simplified, abstracted
Portal Apps Apps Apps network API
Open, programmable API
SDN Controller
(multi-vendor and multi-technology)

NETCONF/
YANG

Disaggregated OLT System


Nx100G Uplinks
Biz ToR/Leaf Switch
ONT ToR/Leaf Switch
TWDM

NGPON2 OLT
MDU
ONT CEx WM1 NGPON2 OLT
NGPON2 OLT
NGPON2 OLT
SFU
ONT
GPON Expandable
General Business
265
ECO System market update
Challenges and Solutions for an
NG-PON2-based Fronthaul
The New Radio Access Architecture (New RAN), as defined by
3GPP in TR 38.801, features the disaggregation of the next
generation node B (gNB) into a Distributed Unit (DU) and
Central Unit (CU). The distribution of functionalities between
Luca Valcarenghi CU and DU impacts the characteristic, in terms of capacity
and latency, that the CU-DU connection (i.e., the fronthaul)
Scuola Superiore Sant’Anna,
shall guarantee. This presentation shades some lights on the
Pisa, Italy
suitability and on the attentions to be taken when an NG-
PON2 is utilized to carry fronthaul traffic.
266
Challenges and Solutions for an
NG-PON2-based Fronthaul

L. Valcarenghi

Scuola Superiore Sant’Anna, Pisa, Italy

NG-PON2 COUNCIL Workshop


“NG-PON2 REALITY NOW!”
th
March 15 , 2018, OFC San Diego, Convention center
Summary

• Ingredients
• Objective: “the cake”
• Recepy
• Experiments
• Conclusions

© 2018 Scuola Superiore Sant’Anna


268
Virtualized New RAN

• gNB functional split • RAN split


– Distributed Unit (DU) – Fronthaul
– Central Unit (CU) – Backhaul

Micro-cloud/fog node

Cloud node
Fronthaul
Backhaul

gNB
© 2018 Scuola Superiore Sant’Anna
269
gNB Functional Splits (3GPP TR 38.801)

codewords layers antenna ports

Modulation Resource element OFDM signal


Scrambling mapper
mapper generation
Layer
Precoding
mapper
Modulation Resource element OFDM signal
Scrambling mapper
mapper generation

Option 7-1 Option 8

© 2018 Scuola Superiore Sant’Anna


270
Functional split in NR L1 (3GPP TR 38.816 V15.0.0 (2017-12))
MAC

Option.6
L1

Coding De-coding

Rate matching Rate de-matching

Scrambling De-scrambling
Option.7-3
(DL only)
Modulation De-modulation

Layer mapping IDFT

Channel estimation
/Equalization
Option.7-2
Pre-coding

RE mapping RE de-mapping
Option.7-1

IFFT/CP addition FFT/CP removal

RF
Digital to Analog Analog to Digital

Analog BF Analog BF

© 2018 Scuola Superiore Sant’Anna


271
Fronthaul requirements (TR 38.801)

Not yet
clarified

ASSUMPTIONS

To be reviewed due to 4ms


HARQ process

Source: LTE for UMTS: Evolution to LTE-Advanced, 2nd Edition


Harri Holma, Antti Toskala
ISBN: 978-0-470-66000-3 © 2018 Scuola Superiore Sant’Anna
272
NG-PON2

© 2018 Scuola Superiore Sant’Anna


273
Ongoing evaluation: simulation
Centralized Coordinated Scheduling (CCS) Distributed Coordinated Scheduling (DCS)

© 2018 Scuola Superiore Sant’Anna


274
Ongoing evaluation: experiment in Advanced
Research on Networking (ARNO) Testbed

• Openairinterface-based experiment
Switch

ONU
OLT

Switch Switch

© 2018 Scuola Superiore Sant’Anna


275
Conclusions

• Overview of fronthaul ingredients


• Overview of fronthaul capacity and latency requirements
• Ongoing evaluation
• Simulations
• Experiments
• If required capacity is small (close to backhaul)
• XGS PONs can be used for lower layer functional split
• … but latency shall be low
• If required capacity is large NG-PON2 or dedicated
wavelength shall be used

© 2018 Scuola Superiore Sant’Anna


276
Thanks

• K. Kondepu
• A. Marotta
• F. Giannone
• P. Castoldi

© 2018 Scuola Superiore Sant’Anna


277
thank you!
email:
[email protected]

This work has been partially funded by the EU H2020


5G-Transformer Project (grant no. 761536)

© 2018 Scuola Superiore Sant’Anna


278
ECO System market update
The Effect of Virtualization on PON
Architectures and Hardware
Today’s PON systems are massively integrated and
application specific systems that, in addition to the PON
functions, contain multiple layers of switching and traffic
management. PON systems are becoming virtualized and the
Ed Boyd impact on system hardware is subtle, yet significant. In this
CTO and Co-Founder, presentation we will compare the integrated and virtualized
Tibit Communications architectures and explore the advantages of virtualized PON.

279
The Effect of Virtualization on PON Architectures
and Hardware
Ed Boyd, CTO & Co-Founder Tibit Communications

280
© TIBIT COMMUNICATIONS, INC., PROPRIETARY
Virtualization Objectives

• Management and Control Application-specific HW


planes Mgmt. Control
Data Plane
Plane Plane
– Disaggregate from Data Plane &
centralize
– Unify across access types

• Minimize application-specific
HW in networks
• Deploy common equipment Standards-based HW
• Facilitate multi-vendor CPE Mgmt.
Plane
Control
Plane
Data Plane
interoperability in remote, virtualization environment

281
© TIBIT COMMUNICATIONS, INC., PROPRIETARY
PON in Virtualized Network
Multiple options for component disaggregation:

Ethernet
PON OLT:
Platform Traffic
PON MAC PON Optics
CPU Management Switching

Generic HW PON Specific HW


• Virtualization allows for component disaggregation
• An ideal split would isolate Generic Hardware and PON Specific Hardware on the
smallest granularity possible
• Line Card Modularity and Port Modularity will be considered
– Does either approach really reduce hardware cost?
– Are there significant architectural advantages?
282
© TIBIT COMMUNICATIONS, INC., PROPRIETARY
Driving Modularity into PON
Starting point: Monolithic architecture
PON OPTICS
PON OPTICS
PON OPTICS
OLT

Legacy OLT Chassis: Control Embedded Platform PON PON


SW Switches CPUs MACs Optics

PON-specific,
fixed
subsystem

component:

modular
component: subsystem

283
© TIBIT COMMUNICATIONS, INC., PROPRIETARY
Driving Modularity into PON
Mid-point: Line-card modularity

Legacy OLT Chassis:

OLT
Control Embedded Platform PON PON
SW Switches CPUs MACs Optics

PON OPTICS
PON OPTICS
1U Pizza Box OLT: PON OPTICS

OLT

Control SW ToR/Agg Embedded Platform PON PON


Control SW Switch CPU MACs
(Remote) Switch Optics

optional

284
© TIBIT COMMUNICATIONS, INC., PROPRIETARY
Driving Modularity to the Port OLT
End point: Fully modular OLT
Control Embedded Platform PON PON
Legacy OLT Chassis: SW Switches CPUs MACs Optics

Pizza Box OLT:

OLT
Control SW ToR/Agg Embedded Platform PON PON
Modular OLT:
(Remote) Switch Switch CPU MACs Optics

optional OLT
Control SW Any PON PON
(Remote) Switch MAC Optics

mandatory

285
© TIBIT COMMUNICATIONS, INC., PROPRIETARY
Anatomy of a 1U Pizza Box OLT PON-specific components

standard Fixed component

PON OPTICS
XFP Socket
PDU 10G PON
MAC XFP Socket PON OPTICS
ToR / Aggregation Switch

XFP Socket PON OPTICS

Fans
DRAM DRAM 10G PON XFP Socket PON OPTICS
MAC
DRAM DRAM CPU XFP Socket PON OPTICS
Fans
10G PON XFP Socket
TM
DRAM FLASH MAC
Switch XFP Socket
Fans
XFP Socket
DRAM DRAM 10G PON
QSFP
DRAM DRAM MAC 40/100G Uplink Optics
PDU
QSFP
40/100G Uplink Optics

286
© TIBIT COMMUNICATIONS, INC., PROPRIETARY
Anatomy of a Modular OLT

per-port modularity
standard

PON PON
MAC Optics
ToR / Aggregation Switch

No PON specific line card

CORD/NFV/Data Center Switches


• Lower cost because of higher volume and more
suppliers
• Evolving much faster than PON OLTs so cost will go
down with more features

287
© TIBIT COMMUNICATIONS, INC., PROPRIETARY
Implementation Flexibility Switch size fits required density & environment
4-port
Remote
Cabinet
12-port Small CO /
Remote

24-port
CO /
Headend
48-port

8-port Outdoor
hardened Remote Node

288
© TIBIT COMMUNICATIONS, INC., PROPRIETARY
High Density Central Office Aggregation
Optional ToR or
EoR aggregation
Assumptions: Host Switch1
Tibit MP OLTs 。。。
(40 per Switch)
• 20 x Host Switches

。。。。。。。。。。
40 MicroPlug OLTs each
• 1:64 subscriber split

Single-rack Results:
• 10G OLT ports: 800
• Subscribers: 51,000
• Power Consumptions: 10kW
(switches + OLTs)

Host Switch20
。。。

Virtualized Modular OLT reduces CAPEX & OPEX (Power/Space)


289
© TIBIT COMMUNICATIONS, INC., PROPRIETARY
Enabling low-cost, low-power Remote PONs
Layer 2
Trunk optics OLT MicroPlugs
(WDM or PON) Switch
Head End & PON
• Splitting the PON MAC and virtualization ONU
environment
optics from the rest of the OLT
ONU
functions allows for remotely
located PON Terminations ONU

(remote PONs) ONU


trunk fiber
• Extend PON to customers
100km
beyond the reach or fiber ONU

capacity of current OLTs


ONU
• An SFP transceiver with an
integrated OLT MAC allows for
small, modular, and low power PON Mgmt & PON PON
Fiber Node solutions. Control SW MAC Optics

290
© TIBIT COMMUNICATIONS, INC., PROPRIETARY
Port-specific expansion

4G/5G Macro eNB

Small Cell
Backhaul:
(10G PON)

Macro Base Station:


(Ethernet)

Cell Site Router/Switch

291
© TIBIT COMMUNICATIONS, INC., PROPRIETARY
Multi-Service Architecture

NG-PON2 OLT MicroPlug


NG-PON2 ONU
10G Ethernet switch

NG-PON2 OLT MicroPlug

NG-PON2 OLT MicroPlug


1598/1234 1577/1270
10G XGS-PON or 10G EPON ONU

10G EPON or XGS OLT MicroPlug

P2P WDM Ethernet


P2P WDM Ethernet

Head-end or Central Office ODN with colorless splitters

• An Ethernet Switch provides flexibility when selecting services for a point-to-point or PON subscriber
• Operators can easily upgrade services by simply swapping modules in Ethernet switches
• Point-to-Point Optics, EPON, XGS, and NG-PON2 on wavelength basis can be mixed as needed

292
© TIBIT COMMUNICATIONS, INC., PROPRIETARY
Interoperability Solution
vOLT-HA [virtual OLT – HW Abstraction] BAA [BB Access Abstraction]

Virtualization
environment

Multi-vendor
adapter/plug-ins

Open Source software solutions provide interfaces for new vendors and new devices
293
© TIBIT COMMUNICATIONS, INC., PROPRIETARY
Conclusions

• Virtualization enables network disaggregation...

...which enables more smaller, modular architectures

• Dividends of this virtualization manifest in CAPEX & OPEX savings:


– Reduced PON-specific components in network (down to size of single port-pluggable optics)
– Reduced PON-specific costs
– Reduced power consumption
– Leverage standard switches
– Enable port-specific buildout and upgrade
– Facilitate OLT-ONU interoperability

294
© TIBIT COMMUNICATIONS, INC., PROPRIETARY
Thank You.

295
© TIBIT COMMUNICATIONS, INC., PROPRIETARY
Thank you
More at broadband-forum.org

29

You might also like