Autumn 2021 - Optical Connections Magazine

Bringing the World the Latest in Optical Communications News

ISSUE 25 | Q3 2021

MEETING THE MDU CHALLENGE Alternative approaches | p24

HOLLOWCORE CABLE A fibre revolution? | p8

5G FRONTHAUL Meeting the challenge | p14

ECOC 2021 PREVIEW Live in Bordeaux | p30


Our InP-based coherent pluggable transceivers enable higher density with superior performance and proven reliability.

The Lumentum 400G CFP2-DCO pluggable transceiver is a CFP2 MSA-compliant digital coherent optical module designed for terrestrial applications including long haul, metro, and access. The 400G CFP2-DCO module is designed to support all interoperable industry initiatives (OpenROADM, OpenZR+, and OIF 400G ZR) as well as proprietary modes for best performance.

The ex-coherent module is perfect for Ethernet, OTN, and FlexO clients, both NRZ and PAM4, and is compatible with legacy and state-of-the-art platforms designed to take full advantage of the 400G DCO. The 400G CFP2-DCO joins Lumentum’s leading-edge coherent product portfolio which includes our 200G CFP2-ACO and 200G CFP2-DCO.

For more information, please visit the Lumentum website.



BACK TO THE FUTURE Welcome to the autumn edition of Optical Connections and to the return of ECOC in its physical form. Both the magazine and the expo have plenty to offer, reflecting the very latest developments and trends in the fibre optic communications industry. In this issue, we’re taking a deep dive into optical networking technologies as we talk to Lumenisity’s VP, Marketing and Sales, Tony Pearson, about the company’s hollow core cable, which has the potential to massively increase transmission speeds. Lumentum’s CTO office’s David Lewis, takes an in-depth look at the connectivity challenges operators face when designing 5G fronthaul to handle the expected high throughput that the technology will require. We also talk to Nokia’s marketing chief Nick Cadwgan about the transformational changes taking place in both the mobile networking layer across the radio access network and the mobile transport core. Veteran journalist John Williamson examines the numerous attractions for optical network operators and service providers offered by Reconfigurable Optical Add Drop Multiplexers (ROADMs) and regular contributor Antony Savvas looks at the challenges faced by network operators when considering which aspects of network management can easily be automated. Meanwhile, at the customer end of the network, we talk to Emtelle and InCoax about two possible alternatives for easing the pain of providing fibre broadband services to customers living in Multiple Dwelling Units (MDUs). There’s also a full preview of ECOC 2021 in this issue along with selected presentations for the ever-popular Market Focus sessions and a full list of categories for the ECOC Exhibition Industry Awards, the winners of which will be announced and the presentations made in the Product Focus Theatre on Tuesday 14th September, at 3pm. New for this year’s ECOC is the Focus Media Zone, which features a theatre where visitors can watch webinars and presentations and back by popular demand are three FTTx Zones showcasing a range of fibre- related products. But that’s not all we’re excited about! The new Optical Connections webinars have been a great success this year and on page 28, you can find out about previous sessions, how to watch any you might have missed, and which subjects will be covered in the future. Also new is the Telecommunications, Optics and Photonics Conference, which will be held in conjunction with Optical Connections, in Central London on 14-15 February 2022. It will cover topics including telecoms, data centres, photonics and free space optics. Check out our website and emails for further details, including the call for papers and registration. The team at Optical Connections and Nexus Media Events look forward to seeing you in Bordeaux in September and wish you a safe and productive ECOC 2021.


Industry News

8 Hollowcore Fibre Tony Pearson 10 Optical Network Management Antony Savvas 14 5G Fronthaul Challenges David Lewis 16 Keeping the Show on the ROADM John Williamson 20 Evolving the Transport Layer Peter Dykes 22 Cleaning IoT Networks Jay Tourigny 24 The MDU Challenge Peter Dykes 28 Optical Connections Webinars 30 ECOC 2021 Preview 32 ECOC 2021 Market Focus 34 ECOC 2021 Awards 36 Product News

Peter Dykes Contributing Editor




Optical Connections is published by NEXUS MEDIA EVENTS LTD Suite 3, Building 30, Churchill Square, Kings Hill, West Malling, Kent ME19 4YU t: +44 (0) 1732 752 125 f: +44 (0) 1732 752 130


ISSUE 25 | Q3 2021


Dell’Oro Optimistic for Broadband Access Growth

total ONT units consumed on an annual basis. The Chinese FTTH market has matured, with broadband penetration in the country reportedly nearing 80%. Though subscriber growth is slowing, there is still a tremendous installed base of subscribers that will continue to require new ONTs. However, additional growth is expected from the rest of the world— particularly North America and Western Europe. In Western Europe, major operators including Orange, DT, BT OpenReach, and Proximus are all expanding their fibre rollouts and even moving quickly to XGS- PON for symmetric 10G services.

Market analyst Dell’Oro says global spending on broadband access equipment and CPE is expected to have a 3% CAGR from 2020-2025, which is a solid increase from a 0% CAGR in its Broadband Access and Home Networking 5-Year Forecast Report. It says the combination of continued residential subscriber growth and increased capacity utilisation rates noted by global broadband providers will more than offset the negative impacts of component shortages and labour limitations. The spending slump Dell’Oro had expected to see in 2021 after the increased investment levels

penetration rates and the total addressable market for broadband service providers have expanded significantly over the last year. Subscriber growth has also resulted in improved revenue and gross margins for service providers. As a result, providers are pulling forward some of their upgrade projects, including those involving the transition from copper to fibre. Dell’Oro’s five-year CAGR for PON equipment has been increased yet again to 5% from 3%. It finds that China, which has historically accounted for anywhere from 65-80% of total PON spending, has peaked in terms of

of 2020 is not going to materialise. In fact, says the analyst, spending will continue to grow this year as operators deal with continued subscriber additions, as well as competitors increasing their investments in fibre, HFC, and fixed wireless networks. It notes that broadband subscriber additions continue to grow at a furious pace around the world, as the Delta variant continues to limit the return of employees to their offices and has extended hybrid learning environments for students. Even if subscriber growth does slow later this year, it says, broadband

DELTA Fiber Taps Nokia For 400G+ Transport Network

Furukawa, Fujitsu, Collaborate on Next- Gen Integrated Optics

Furukawa Electric and Fujitsu Optical Components (FOC) have agreed to collaborate on product development of integrated devices for next- generation, high-capacity optical communications. The companies say they will utilise the strengths of both companies to develop high-capacity, compact, low power consumption devices for next-generation communications networks, meeting the need for solutions in the Asia region. The aim is to create world-class, top-performing devices for next-generation communications networks by combining the technologies of both companies in order to respond to the explosive increase in communications traffic and demand for reduced power consumption. Furukawa Electric and FOC also aim to bring together the two companies’ optical

device products for digital coherent systems in order to offer specialised transceiver solutions to customers in the Asia region, a market where such solutions will be in high demand. Furukawa says the collaboration will bring together Furukawa Electric’s compound optical semiconductor technology and FOC’s LN/silicon photonics technology to develop integrated devices

Nokia says it has been selected by Netherlands operator DELTA Fiber to provide a next- generation optical transport network, based on 400G wavelengths, to handle the operator’s increased traffic and further expansion of its FTTH rollout. The high-capacity network will offer customers enhanced service quality and speeds. It adds that 400Gbps speeds and higher wavelengths enable a simplified network that increases operational and cost efficiency. To rapidly deploy services to its customers, reduce network total cost of ownership and extend network lifecycles, DELTA Fiber will use the

Nokia 1830 Photonic Service Switch (PSS) platforms. This, says Nokia, will support DELTA Fiber’s deployment of a new DWDM network, incorporating Nokia’s broad family of ROADMs, enabling optimized core and metro applications to cover the entire country. The core network build is currently underway to support 19 sites and will be followed by the deployment of metro sites, covering approximately 75 locations. This deal is part of a broader cooperation with DELTA Fiber to support its expansion plans, which also includes Nokia supplying XGS. PON access network and customer premise equipment.

for next-generation, high-capacity optical

communication that are both high in performance and compact in size, bringing together both companies’ technologies to create something that cannot be achieved by these components individually. The companies say they intend to deploy these integrated devices globally for the over 800Gbps transceiver market.


| ISSUE 25 | Q3 2021


EXFO Shareholders Approve Go-Private Arrangement

DZS Expands Collaboration with Broadcom

Optical, mobile and broadband solutions

EXFO says that, at the special meeting of its shareholders held on 13th August, a significant majority of shareholders voted in favour of the special resolution approving the previously announced Go-Private plan under Section 192 of the Canada Business Corporations Act pursuant to which 11172239 Canada Inc. will acquire all the issued and outstanding subordinate voting shares of EXFO, other than the subordinate voting shares held by founder and executive chairman Germain Lamonde, for US$6.25 per subordinate voting share in cash. Shareholders carrying an

aggregate of 333,666,141 votes, representing approximately 97.49% of votes entitled to be cast at the Meeting, were represented in person or by proxy at the meeting. The Arrangement Resolution was approved by 99.65%

centralised or distributed networks with state-of-the- art orchestration and network automation for software- centric services and network virtualisation. Broadcom’s broadband access technology currently underpins DZS Velocity solutions, including the recently announced XCelerate by DZS XGS-PON solutions, as well as DZS Helix connected premises solutions in ONT fibre termination point and home Wi-Fi products. At the same time as the announcement, DZS said Misty Kawecki has been appointed as chief financial officer, effective immediately. She succeeds Tom Cancro, who served as CFO since 2019.

provider DZS has expanded collaboration with Broadcom to advance open standards and bring to market next- generation high-speed optical broadband and PON technology. DZS says Broadcom’s scalable semiconductor devices provides performance, capacity and flexibility for many DZS optical and copper-based products. Integrated with the DZS SDN-enabled Network Operating System (sdNOS) for network-based products, and supported by the DZS Cloud Portfolio, Broadcom technology enables DZS’ service provider customers to implement ultra-performance

of the votes cast by shareholders, voting

together as a single class, as well as 90.95% of the votes cast by the holders of subordinate voting shares, excluding the votes attached to the Excluded Shares. The Arrangement remains subject to customary closing conditions including the approval of the Superior Court of Québec.


ISSUE 25 | Q3 2021


Orange Polska Tests Seven-core Fibre Reaching 11.2 Tbps

Prysmian Group Extends Cable Range for FTTx and 5G Networks

Orange Polska’s Innovation Lab, together with Infinera and InPhoTech group, tested multi-core fibre and Infinera’s ICE6 800G technology. Infinera claims the throughput obtained during the tests was seven times greater than the maximum that can be achieved today using a standard fibre optic cable. Developed within the

Prysmian Group has extended its Sirocco HD range of microduct cables to include a 576 fibre optic cable in a diameter of 9.5mm, providing a fibre density of 8.1 fibres per mm2, and is installable into a 12mm duct. Sirocco HD microduct cables utilize Prysmian’s BendBright-A2 200µm single-mode (ITU-T G.657.D, G.657. A2) bend insensitive fibre, providing a solution that’s ready for evolved systems and is future-proof. Available now in fibre counts from 96 to 576 and conforming to international standards for

simultaneously. This means that its capacity is seven times greater than that of a standard

optical and mechanical performance, the Sirocco HD cables also benefit by the use of Prysmian’s PicoTube technology making them up to 20% smaller than previously available microduct cables. This makes it possible to install more fibres into congested duct space, and enables the use of smaller ducts for new installations, resulting in lower installation costs and the use of less raw materials. This provides benefits for both the total cost of network deployment and the environmental footprint.

telecommunications fibre. Such optical fibres will be produced in Lubartów by IPT Fiber from InPhoTech group. Infinera says tests have shown that the its hardware allows for 800Gbps data transmission in a single transmission channel. The experiment, conducted in collaboration with Orange, used two channels sending data at a speed of 1.6Tbps in each of the seven cores simultaneously. This gave a total transmission of 11.2Tbps.

InPhoTech group in cooperation with the

Maria Curie-Skłodowska University in Lublin and with the support of the Photonics and Fibre Optics Cluster, the multi-core fibre allows transmission in seven parallel cores

Nexera Gets EU Funding for Rural Fibre Rollout

on the Polish market, confirming the wide development prospects. “Our success is supported by a well- developed business plan, a real and effective company strategy as well as an innovative way of operating our organization, based on the “agile” model, which has in many respects become a benchmark for other operators in the market,” said Jacek Wiśniewski, president of the Nexera Management Board. “We are a relatively small organization, while

in Nexera Regions. The list of lenders includes The European Investment Bank, PKO BP Bank, Santander, BGK and ING. The funds obtained will allow Nexera to pass more than 700,000 addresses that will ultimately be within the range of the operator’s fibre-optic network. In June of this year, Nexera purchased fibre-optic infrastructure from IT Partners Telco, operating locally under the Navu brand. The acquired network consists of more than 5,000 households located in Sandomierz.

Polish operator Nexera received over US$266 million in financing from a consortium of banks including around US$85 million from the European Investment Bank for the development of fibre- optic networks in rural Poland. The funding, in the form of a loan, will be allocated to investments in the development of FTTH networks in the regions. When comparing the size of the company to the amount of the funds made available, Nexera says it is an unprecedented event

the debt financing we receive places us at the level of the largest players in the country. This is a clear success in the Polish market, which shows that we are a reliable organization with huge potential, as confirmed by leading financial institutions,” said Paweł Hordyński, Member of Nexera’s Financial Management Board. The financing will be provided in the form of an overdraft facility and an investment loan to be used for the development of fibre-optic infrastructure


| ISSUE 25 | Q3 2021


• Very small form factor muti-fiber connector • 3x density over MPO connectors • 16, 8 and 4 fiber TMT ferrule technology • Standard 250 µm fiber pitch • IEC grade B performance • Single-mode APC and multi-mode APC • DirectConec ™ push-pull system

High Density Connectivity Revolution Unfolds




The capabilities of Lumenisity’s NANF® hollowcore fibre have caused something of a stir in the fibre optics communications industry. Optical Connections editor Peter Dykes spoke to Tony Pearson , Lumenisity’s VP Sales & Marketing, to find out more about this potentially game-changing technology.

How did the project to develop NANF hollowcore fibre come about and why was it decided to

non-linearity and lower dispersion and is able to handle higher power levels than conventional single-mode fibre (SMF). These performance benefits can be exploited to help overcome the capacity and reach challenges in today’s optical networks. We have made CoreSmart interfaces compatible with SMF and associated amplifiers and transmission equipment, this extends the total reach and the number of possible future applications.

can be fusion spliced using certain types of commercially available splice machines with a purpose designed fusion splice recipe. This allows field jointing of multi-cable segments in challenging outdoor applications. It can also be jointed to single-mode equipment such as termination hardware, optical multiplexers and amplifiers, greatly extending the reach and range of networking applications compared to other conventional legacy hollowcore fibres, for example 5G and DCI. Lumenisity’s customers recognise the performance advantages of CoreSmart that help to drive greater value in the products and services they can offer their customers, however as a policy, Lumenisity does not publicise any cost or pricing information.


create a spin off company?

Lumenisity was formed in 2017 as a spin-out from the University of Southampton to commercialise


breakthroughs in the development of hollowcore optical fibre. The initial investment was in high-frequency trading (HFT) as a clear use case. Now the technology has developed further, we are expanding our focus markets. Lumenisity first deployed hollowcore cable carrying production traffic in February 2017 and has cables deployed in diverse geographic locations carrying traffic in indoor and challenging outdoor environments, and we’ve been granted many patents that cover NANF hollowcore fibre and interconnect technologies.

Is it more difficult to manufacture than conventional fibre?


There are many different sub-categories of ‘conventional fibre’ that operate in different


applications. The solid-core fibre industry has had more than 40 years to evolve, whereas NANF technology is in its early infancy, and its full potential is yet to be exploited as we continue to make performance improvements.

Is hollowcore fibre suitable for all fibre communications applications, or is it suited to


more specific applications? If the latter, why is it so suited and what distances can it carry data?

What are the advantages over conventional fibre?


How does it compare to conventional fibre in terms of characteristics such as


In a conventional optical fibre, light signals travel through solid glass, whereas the light inside a


We are still exploring the range of usage cases where this type of technology can provide most


bendability, ease of use and cost?

hollowcore fibre is guided along a core structure that comprises of air. Since light travels more quickly in air, data signals travel 50% faster in Lumenisity’s NANF CoreSmart® cable. CoreSmart has lower

benefit, especially as attenuation is improved. In summary, in any application there is a 2.25 increase in geographic footprint for any given latency envelope:

TP CoreSmart can also be deployed in indoor and outdoor situations use the ducting, racking and patch panels just like conventional SMF. NANF


| ISSUE 25 | Q3 2021


Does it require non-standard connectors e.g., for connecting to transceivers, patch panels etc

dispersion and can be operated over a broader wavelength range e.g., O, C & L bands. This increases system reach and data carrying capacity. Also, CoreSmart cables are field-splice capable allowing multiple cable segments to be jointed and deployed in challenging outdoor environments to service longer route length networking applications. In addition, they are compatible with SMF and equipment, such as optical amplifiers, this greatly extends the reach and number of possible network applications. The performance of our hollowcore technologies has been evaluated by independent parties, including carriers and equipment providers and typically shows 100x longer reach with 40x the data-rate over other types of hollowcore fibre.

this is true in all applications. This saves overall cost and gives flexibility in access, for example to power utility. Multi- channel, multi-terabit transport has been demonstrated over several thousands of kilometres using commercially available transmission equipment. We expect significant advantages for applications such as HFT, 5G networks and enabling increased geographic expansion in DCIs, that benefit from the much lower latency than solid core fibres and other types of hollowcore that are route distance and bandwidth limited. There have been several successful transmission trials with CoreSmart hollowcore fibre. BT labs showed error-free transmission with 38 DWDM channels over 10km of field-deployable CoreSmart cable using 400ZR pluggable transceivers. Lumenisity and Ciena completed DCI transmission trials using Ciena’s WaveLogic 5 Extreme (WL5e) and Wavelogic 5 Nano (WL5n) commercial systems and achieved a capacity of 38.4Tbps with 48 x 800Gbps channels over greater than 20.5km of CoreSmart cable, without in line amplification. Characterization testing was also carried out on CoreSmart cable in a recirculating loop, achieving over 1000km reach across the C band with 45x 400Gbps WL5e channels.


TP CoreSmart solutions can be jointed to conventional SMF, so existing termination equipment such as transceivers and connectors can also be used.

Can hollowcore fibre be used in conjunction with conventional fibre on the same network? CoreSmart cable can be used in conjunction with conventional SMF over the same network. It is



fully backward compatible with existing networks and systems and can be fusion spliced in the field and can be jointed to other types of SMF equipment, such as optical amplifiers, splitters and multiplexers.

What is the roadmap for further development?


We are constructing our own dedicated fibre fabrication facility in Romsey, Hampshire, UK. CoreSmart


Other companies have been developing hollowcore fibre, so what is different about


cable has the potential to re-write the network design rule-book. We are developing future generations of CoreSmart solutions with lower attenuation to extend the reach and increase the capacity beyond that of conventional solid core fibre optic cable today.

Lumenisity’s offering?

TP Compared to other hollowcore fibre, such as Photonic Band Gap Fibre (PBGF), our NANF based CoreSmart has lower attenuation, lower


ISSUE 25 | Q3 2021



Efficient optical/fibre network management is a cornerstone for any successful service roll-out, Antony Savvas looks at the challenges and the solutions that the industry must consider.

TRAFFIC MANAGEMENT The issues around traffic management are changing as the connectivity market changes. Wim te Niet, EMEA vice president of sales at EXFO, says, “For many communications service providers (CSPs), the drive to roll out FTTx (home, business and antenna) means optical and fibre connections are being deployed in series alongside existing technologies such as microwave. “This introduces complexities around traffic management given that CSPs often do not have a full, end-to-end network topology view. This means they lack visibility to the consequences of combining multiple transport technologies in a hybrid environment.” Guillaume Crenn, optical transport product line manager at Ekinops, says of traffic management, “One of the most significant issues surrounding traffic/ network management is the growth in channel capacity, which in the last 5-10 years has gone from 10G to 600G, and we are now looking at 800G and 1Tb. While this has made it easier from an optical standpoint, as fewer wavelengths are needed, it’s a challenge for management as services can be aggregated into a single wavelength - often x100 more - making individual service management complex.” From an operator perspective, says

FAULT MANAGEMENT In the same vein, what about fault management, is it becoming more difficult to tackle? Scott Larrigan, product marketing manager for optical networks at Nokia, says, “When we embed optical network expertise and the latest data science into easy-to-use network applications, we can make fault management much easier. Using these capabilities we can monitor network KPIs and trigger proactive network automation to minimise the impact of network faults. “We can also automatically trigger applications to rapidly isolate the cause of faults. For example, using OTDR (optical time-domain reflectometer) tools, we can rapidly find the physical location of a fibre fault,” says Larrigan. On fault management, Ekinops’ Crenn says, “Compared to the 10G capacity lines containing 40 channels, fault management on today’s 600G lines aggregating potentially hundreds of services and multiple standards is a significant challenge. Network operators need to be able to do very granular performance monitoring (PM) in order to provide an end-to-end service level agreement (SLA) on each service individually.” This requires the ability to capture and analyse the PM counters

Crenn, managing multiple services on a high-capacity line can be a significant challenge. “For Tier 2-3 service providers, achieving the balancing act between short-term capex and long- term opex is difficult to assess,” he says. There is a temptation to turn to a less expensive fixed-grid architecture when scoping for a network upgrade, but this limits scalability in both the number and line rate of the channels that can be added. For instance, says Crenn, an 800G channel doesn’t fit over a 100GHz fixed filter network. A more costly flex-grid solution is more future-proof however, and will ultimately result in lower opex over time, but it is “harder to quantify when they are developing the business case”, he adds. Jonathan Homa, senior director of portfolio marketing at Ribbon, adds, “To improve resilience while optimising overall capex, traffic management must coordinate effectively between the IP and optical layers. This is driving a trend to multi-layer optimisation (MLO) where the management tools for each layer are aware of the other connectivity layers. As an example from the perspective of the optical layer, when it needs to perform link restoration (e.g. when using wavelength switched optical network, or WSON) it should give preference to links with higher service priorities.”

in order to identify events such as bit errors and loss of frame, and


| ISSUE 25 | Q3 2021


CAPACITY OPTIMISATION And how about capacity optimisation, are organisations successfully achieving it or are they simply throwing in more bandwidth to satisfy their needs? Nokia’s Larrigan says on this, “Optical network operators face growing capacity demands and the cost of addressing these requirements can significantly impact their bottom line. With the introduction of SDN (software-defined networking) optical networks will see increased client port speeds and more dynamic service requests. “For network operators, the challenge is to extract maximum efficiency and bridge the gap between where their networks are today and where they need to be in the future.” Advances in coherent wavelength modulation formats, together with Colorless Directionless Contentionless - Flexgrid (CDC-F) wavelength routing, are enabling new approaches to optical network optimisation, says Larrigan. “Rather than defining beginning-of-life network infrastructure based on worst- case, end-of-life fibre infrastructure parameters, operators can use real-time network data within a continuous ‘learn and optimise’ approach that adapts to approaching end-of-life conditions,” he says.

The idea is that operators can use applications to periodically tune the network to maintain optimal performance and availability and stay ahead of deteriorating network conditions. When end-of-life conditions eventually threaten network performance, the application can make proactive network re-optimisation recommendations before they impact service deployment velocity, resulting in a more robust and adaptable network. Homa from Ribbon said, “The benefits of MLO have always been clear, with the potential to produce five-year capex savings from 15% to 60% depending on how deeply it uses shared resources for restoration. But until recently this has been difficult to implement practically because of the algorithmic and control complexity. What is happening now is a two-pronged approach. Organisations are continuing to increase bandwidth capacity since new technologies like 400G ZR+ allow them to do so economically, and they are also beginning to adopt aspects of MLO because new SDN tools are now making this possible,” said Homa. AUTOMATION Automation is arguably key, are there serious attempts in the market already to address the above issues,

to correlate alarms at multiple levels - service, wavelength, link etc. - to effectively identify the root cause. “The way to accommodate PM in this increasingly complex environment is the NMS (network management system) software, which needs to not only be dimensioned to handle more services, but also to provide an intuitive user-interface to enable simplified fault

management,” Crenn added. Ribbon’s Homa concluded,

“Enigmatically, fault management is becoming both more complex and easier to execute. Complexity is added if we want to consider the full impact of a fault on all layers of the network, not just an isolated, single optical layer or IP layer. But today’s advanced software tools automate many of the processes involved with fault management such as fault identification and localisation, service and network restoration for increased ease.” Homa said, “Multi-layer optimisation is an important tool in the restoration toolkit, but the network must be designed in such a way that it doesn’t end up with fault restoration processes in the IP layers being in a race with restoration processes in the optical layer - that can easily cause network instability.”


RWC12038_Telecoms_Optical_Connections_HDPS_120x420_AW.indd 1


| ISSUE 25 | Q3 2021


perspective and from a network management perspective.” Sjoberg cites PagerDuty for general network monitoring and providing a single place to view the overall health of systems and operations. In addition, Elasticsearch - an open source, full-text search and analysis engine - is used within ELK stacks, and Packetbeat provides insight into how traffic is flowing through networks, he says. “Tools such as these are available to gain observability and secure the environment, ensuring high levels of performance,” said Sjoberg. As the demands on network management evolve, it appears that the industry is incrementally meeting any challenges.

It allows service providers to look at trends, take in multiple data sources and then use machine learning and artificial intelligence to determine likely future events or issues. “This means they can fix potential issues around network capacity, for instance, before customer service is affected - leading to increased customer satisfaction and the reduction in any SLA penalties,” he added. Gustav Sjoberg, head of network operations at digital infrastructure deployment and management specialist VX Fiber, says, “Our senior network engineers and NOC engineers are able to manage multiple full-scale networks with relative ease thanks to automation in various forms, both from a monitoring

or is it still pie in the sky for still too many organisations? EXFO’s Te Niet says service providers are adopting automation to ensure SLAs. He says, “Correlating service-affecting issues to determine which layer has the problem reduces MTTR (mean time to repair) times significantly. Automating anomaly detection allows the service provider to achieve this goal. By correlating different data sources automatically - such as the fibre monitoring, OLT, ONU, L2 and/or L3 service - service providers can quickly find the service-impacting issue and then dispatch to fix, rather than dispatch to find the problem,” said Te Niet. He said predictive analysis is key to preventing customer-impacting issues.

Wim te Niet EMEA Vice President Sales EXFO

Guillaume Crenn Optical Transport Product Line Manager, Ekinops

Jonathan Homa Senior Director of Portfolio Marketing, Ribbon

Scott Larrigan Product Marketing Manager Optical Networks, Nokia

Gustav Sjoberg Head of Network Operations VX Fiber

Super-fast, super-secure connections for communities Discover howour push-fit technology provides the ultimate in fast, strong and reliable blown fibre connections. Our product ranges include Direct Buried, Protected and Gas andWater Block Connectors for fibre cables andmicroducts. With all our fittings you can be reassured that

they are engineered to last, in the toughest environments. Thank you to all our customers for helping us reach our 60th anniversary. Here’s to the future.

04/08/2021 11:21


ISSUE 25 | Q3 2021



To meet the demands created by the expected 3.5 billion 5G users by 2026, carriers need to deploy fronthaul technology that can handle the high throughput of 5G. However, the wide range of applications and use cases have varying requirements, and there are many architectures that can serve each. This makes deciding on a fronthaul approach a challenging prospect because the market has no clear choice on which to converge, writes David Lewis , CTO Office, Lumentum.


and have different reach requirements in the fronthaul network. To address this, multiple WDM maps with multiple RAN partitioning architectures over duplex and bidirectional fibers are being applied. To keep costs down, 5G fronthaul connections are being designed using existing and mature optical technologies. These technologies were originally developed for highly cost- sensitive and high-volume data center

to six wavelengths in each direction. Low channel count and moderate reach WDM connections like these are pervasive in data center networking. In such cases, O-band (1260 to 1360 nm) wavelengths can be reused. A key advantage of this approach is that this reuse leverages the massive investments already made in maturing these O-band wavelengths for hyperscale data center interconnects. The three O-band contenders are coarse-, modified, and LAN- wavelength division multiplexing, also known as CWDM, MWDM, and LWDM respectively. CWDM is based on 20 nm spacing between channels with a 13 nm passband at the receiver. Such wide spacing allows for uncooled operation over a wide temperature range. Mobile network designers have added two wavelengths (1351 and 1371 nm) to the four wavelengths already used in data center networking: (1271, 1291, 1311 and 1331 nm). For architectures with separate up- and down-link fibers, six 25G eCPRI transceivers each operating on one of the CWDM wavelengths can be used together for an aggregate bandwidth of 150 Gb/s in each direction. For architectures with scarce fiber resources, MWDM achieves the same throughput as CWDM on a single

Fronthaul connectivity between active antenna units (AAU) utilizes either a distribution unit (DU) or a central unit (CU) that can operate over a distance up to ~ 10 km (see Figure 1). Spectrum allocations per carrier of 100 to 200 MHz mean that each AAU needs a bidirectional data rate of 25 to 50 Gb/s based on 1 or 2 ports of 25G eCPRI.

Figure 1: 5G fronthaul uses either a distribution unit (DU) or central unit (CU)

networking applications. Thus, 5G fronthaul equipment is being built on proven technologies that have already been cost-reduced through volume production. WAVELENGTH GRID CHOICES FOR 5G FRONTHAUL Many fronthaul

5G deployments aggregate the data from multiple AAUs resulting in 100’s of Gb/s of bidirectional data per DU/ CU over the fronthaul connections. As these fronthaul connections are often fiber constrained, wavelength division multiplexing (WDM) is used because it enables multiple AAUs to share the same fiber by allocating different wavelengths (channels) to the data of each AAU. In addition, the 5G rollout is happening in all the world’s regions, requiring network operators to decide how to architect the radio access network (RAN). Distributed RAN (D-RAN) and centralized RAN (C-RAN) architectures call for different numbers of channels

architectures connect three AAU base stations to a single distributed unit (DU), as shown in Figure 2. Such fronthaul networks can be realized with a small number of WDM channels, such as three

Figure 2: 5G fronthaul with 3 AAUs per DU


| ISSUE 25 | Q3 2021


bidirectional fiber. It does this by transmitting six wavelengths up and six wavelengths down in a bidirectional fashion. The six receivers at each end of the link have the same 13 nm passband as the CWDM grid. The modification made by MWDM is to have two channels, each offset from the standard CWDM channels by ± 3.5 nm. For reaches beyond 5 km, the longer wavelengths of CWDM and MWDM are subject to high chromatic dispersion. Thus, for longer reach systems of 5 to 15 km, it is necessary to move away from wavelengths significantly above the fiber zero dispersion wavelength. This is achieved by using LAN WDM (LWDM), which is considered by the ITU-T to be dense WDM (DWDM) with channels on an 800 GHz spacing. For systems where more channels per fiber are required, the next step up is dense WDM (DWDM) in the C-band (1530 to 1565 nm). With 100 GHz channel spacing, WDM links support up to 40 channels per fiber. This makes large, centralized RAN (CRAN) fronthaul networks in fiber-constrained applications possible. The basic parameters of the four grids are listed in Table 1. Table 2 compares typical implementations of the four grids in terms of the type of transmitter device.

As can be seen, the trade-offs between fiber count, dispersion penalties, and requirements for laser temperature control mean that “one size does NOT fit all”. Given the range of topologies and applications demands, it is likely that

Figure 3: Lumentum DFB laser chip photograph and schematic

TUNABLE SFP+ TECHNOLOGIES FOR 5G FRONTHAUL Lumentum has a long history supplying tunable transceiver modules with reaches up to 80 km (10 Gb/s) or 15 km (25 Gb/s). Industrial temperature versions operate from -40 to 85 °C and have power consumption less than 2.5 W. With over two million tunable XFP and SFP+ modules shipped, Lumentum technology has been proven in the field many times over. These transceivers are built around Lumentum digital supermode distributed- Bragg-reflector (DSDBR) wideband tunable laser. The laser is monolithically integrated with an InP Mach-Zehnder modulator (MZM). An integrated semiconductor optical amplifier (SOA) provides optical power control and acts as a shutter to allow dark tuning when reverse biased. In addition, automatic wavelength tuning allows for software-controlled, full C-band tunability. Thus, any single module can flexibly support any channel. Because the modules are plug-and-play, setup requirements are greatly reduced. There is no need to track fiber and identify the proper add/drop channel on the mux/demux facility. This means a single TSFP module with automatic wavelength tuning replaces a large list of part numbers when fixed DWDM transceivers are used. This provides further operational and supply chain efficiencies with improvements in both OpEx and CapEx for operators. The result is lower module cost, simplified maintenance, and reduced inventory investment and management. CONCLUSION 5G fronthaul networks are being deployed on a variety of WDM wavelength grids. Low channel count networks with 6 or 12 wavelengths are in the O-band and employ three standard approaches. For higher channel count networks, C-band DWDM on 100 GHz using tunable modules is gaining traction. Automatic wavelength tuning modules offer significant OpEx and CapEx savings for mobile operators needing high channel counts.

all grids will co-exist in 5G fronthaul networks. Thus, mobile operators will be evaluating and deploying all of the above approaches for fronthaul connectivity. ENABLING 5G FRONTHAUL WITH DIRECTLY MODULATED LASERS Directly modulated lasers (DMLs) are used in high volume for optical transceiver modules such as SFP28 and QSFP28. In the datacom market, standards such as 25GBASE-LR and 100G-CWDM4 utilize uncooled 25 Gb/s DMLs within the data center environment. These same 25 Gb/s DML chips can be used for 5G fronthaul on CWDM, MWDM and LWDM grids with temperature control in mobile outdoor plant environments. Lumentum is a leading supplier of 25 Gb/s DMLs providing market-leading power efficiency, wide temperature

Table 1: Wavelengths, channels, and spacing for 5G fronthaul grid plans








1271 – 1371

20 NM



1267.5 – 1374.5

7 NM



1269.23 – 1318.35 227.4 – 236.2 800 GHZ



1529.55 – 1560.61

192.1 - 196.0 100 GHZ

Table 2: Comparison between grids for 25 Gb/s transceivers on bidi fiber












COOLED 5 – 10








COOLED 10 - 20



operation, self-hermetic chip designs, and high bandwidth enabled by a corrugation pitch modulated (CPM) design (see Figure 3).

DML: Directly modulated laser, typically a direct feedback laser (DFB) EML: Externally modulated laser, typically a DFB plus electro-absorption (EA) modulator MZM: Mach-Zehnder modulator Uncooled: DFB center wavelength varies by ~0.1 nm/°C. Temperature control keeps each channel within the tolerance of the grid. Max reach: The combination of transmitter chirp and fiber dispersion limits transmission distance for 25 Gb/s DML transmitters at wavelengths above 1330 nm.

Figure 4: Lumentum has a long history supplying tunable transceiver modules


ISSUE 25 | Q3 2021



Reconfigurable Optical Add Drop Multiplexers (ROADMs) have numerous attractions for optical network operators and service providers. So much so, in fact, that an analysis available from the TECHNAVIO market research house is forecasting the value of the global ROADM market could reach US$5 billion by 2024, up from US$2 billion in 2019, writes JohnWilliamson .

E xpert opinion differs somewhat as to which of the various ROADM improvements over OADMs is of most consequence. Helen Xenos, Ciena’s senior director, Portfolio Marketing, opts for the flexibility they deliver, in what is an unpredictable environment, to evolve the network in terms of scale and more easily support new coherent technologies. She also cites re-configurability that enables rerouting of wavelengths as needed and the management of power profiles when moving from one fibre to another. As well as network flexibility, re- configurability and boosting network expansion, Raj Nagarajan, Lumentum’s senior product line manager, mentions optical by-pass of express traffic which helps eliminate optical-electrical-optical regeneration to lower the overall network ownership cost. Elsewhere in the ROADM industry playbook claims are made for improved network utilisation; accelerated service provision; simplified network management; higher network resilience and automated network and node reconfiguration. ROADMWORKS From their initial use to remotely switch traffic in long haul Wavelength Division Multiplex (WDM) networks in the early 2000s, ROADM repertoire, technology

hardware innovation. Given average traffic growth rates of 30% to 50%, she also reckons the addition of L-band is a big deal. “With respect to ROADM technology, the biggest change we have seen is the move to C- and L-band systems,” she says. Migration to L-band has had teething problems with regard to it sometimes requiring a non-trivial manual engineering effort, and there’s the prospect of possible traffic loss through adverse impact on the in-service C-band channels. Xenos says one solution is Amplified Spontaneous Emission (ASE) loading. According to a Ciena account, filling the unused spectrum with ASE maintains full loading conditions on the system, so that system performance can be known and optimised at the time of deployment. As a result no additional margin needs to be budgeted for an L-band upgrade, and system performance will remain consistent across its lifetime, regardless of channel loading conditions. Whether the L-band is carrying live traffic or not, the system behaves as if it is fully loaded from day one. Karl Heeks, senior programme director, ROADM Systems Group, HUBER+SUHNER, describes how increased capacity is being provided by increased bandwidth within the super C and C-& L-bands, broader spectral channels to support higher baud rates, together with spatial parallelism with multiple fibres per degree, and perhaps ultimately, multi-core or multi-mode fibres in some applications.

and areas of application have evolved considerably. On the extended repertoire front, modern ROADMs can offer ‘Colorless’ (C), ‘Directionless’ (D) and ‘Contentionless’ (C) capabilities, collectively styled ‘CDC’. The first enhances the basic ROADM with the ability to connect any channel to any mux port, the second enables the sharing of muxes between directions, and the third enables multiples of the same wavelength to be added/ dropped to the same mux. In their most ambitious iterations, ROADMs add ‘Flexible-grid’ (F) to CDC to provide a degree of future-proofing with the ability to dynamically accommodate elevated traffic demands. Not everyone is in the market for the full-blown CDC/F treatment. Requirements vary from operator to operator, and from network to network. Cost may be an issue. As Nagarajan observes, Colorless is cheaper than Colorless combined with Directionless, which in turn is cheaper than CDC. “We really do not see that a ‘one size fits all’ approach is what our customers are asking for,” adds Xenos. NUTS AND BOLTS CHANGING Meantime, the nuts and bolts of ROADM technology is changing. Here Xenos includes an increased adoption of compact modular disaggregated line systems, and the circumstance that software has become as important as


| ISSUE 25 | Q3 2021


support the greater number of ROADM degrees without the complexity and cost of passive splitting and parallelised modules.” Heeks reports that a new generation of high-capacity WSS devices in development at HUBER+SUHNER makes use of ultra-high definition 4k LCoS panels and 2D holographic beam- steering. “The expectation is that this disruptive WSS platform will allow the reimagining of ROADM architectures across all levels of the network and pave the way for a new class of ROADM that enables contentionless routing of both WDM and SDM traffic in an integrated network node,” he states. As future prospects Xenos instances: multi-degree integration in a single module; higher integration of modules supporting multi-rail architectures; and more and more embedded instrumentation. “There’ll also be more ‘design-from-the-ground-up with modern container-based software architecture,” she anticipates. But it’s likely that there will be bumps in the ROADM ahead. “There are going to be many challenges to overcome,” concludes Xenos. “For example, how to manage software upgrades? How are we going to connect systems with different WSS technology? Who is going to do the integration testing? And trouble-shooting will be difficult.”

There’s also demand for ROADMs in long haul Data Centre Interconnect (DCI) networks at >100 km, while Lumentum sees a new ROADM application space opening up at the edge of the network to support the increased traffic growth driven by 5G, edge computing and content caching. THE ROADM AHEAD So, what’s coming down the ROADM turnpike? “Going forward”, predicts Nagarajan, “ROADM will continue to evolve to improve the cost per unit bandwidth as network capacity grows, taking advantage of higher levels of WSS integration such as collapsing the functions of separate C- and L-band WSSs, allowing continuous C+L operation in a single device.” Heeks agrees that upcoming ROADMs must not only deliver increased capacity and versatility, but must do so at increasingly lower cost-per-bit. This will require a new generation of WSS, with higher port counts and flexible configurations. “Line-side WSSs will require future-proofing against both degree and bandwidth increases, to allow pay-as-you-grow expansion of the nodal capacity. Spectral coverage is less critical for client-side switches, where the bandwidth can be split across multiple modules,” he maintains. “The key challenge here is, instead, to

“The core network requires high- degree ROADMs - for example an 8-degree node may require a 16-, 24- or 32-degree ROADM for fully CDC operation,” continues Heeks. “Simultaneously the number of client- side ports as a proportion of the ROADM degree is reducing, with fewer but higher-capacity channels to add and drop from each degree.” Other innovations are: the MxN connectionless Wavelength Selective Switch (WSS), in which ‘N’ denotes the number of add-drop ports and M is the number of directions the ROADM supports; and inherent filtering. “The recent introduction of contentionless MxN WSS allows seamless evolution of coherent transmission to higher data rates, enabling lower cost per bit in CDC ROADM nodes, while inherent filtering enhances performance, optimising spectral efficiency x reach,” explains Nagarajan. FORKS IN THE ROADM Nagarajan remarks that today ROADMs are widely used in long-haul and metro networks, with the bulk of the volume deployment being in the latter. “It is worth pointing out that metro volumes are larger simply due to metro networks having a larger volume of network nodes relative to long-haul networks,” he qualifies.


ISSUE 25 | Q3 2021

Page 1 Page 2 Page 3 Page 4 Page 5 Page 6 Page 7 Page 8 Page 9 Page 10 Page 11 Page 12 Page 13 Page 14 Page 15 Page 16 Page 17 Page 18 Page 19 Page 20 Page 21 Page 22 Page 23 Page 24 Page 25 Page 26 Page 27 Page 28 Page 29 Page 30 Page 31 Page 32 Page 33 Page 34 Page 35 Page 36 Page 37 Page 38 Page 39 Page 40 Page 41

Made with FlippingBook - Online Brochure Maker