Relentless traffic growth leads to a ROADM rethink
Technology briefing: ROADMs
Lumentum has developed an optical switch to enable reconfigurable optical add-drop multiplexers (ROADMs) to cope with the traffic growth expected over the next decade.
The company’s MxN wavelength-selective switch (WSS) will replace the existing multicast switch used in colourless, directionless and contentionless ROADMs. The Lumentum TrueFlex 8x24 twin switch will enable networking nodes of 400-terabit capacity.
“This second-generation switch is what will take us into the 100 gigabaud and super-channel era of network scalability,” says Brandon Collings, CTO of Lumentum.
ROADMs
ROADMs sit at the mesh nodes in an optical network. Their function is to pass lightpaths destined for other nodes in the network - referred to as optical bypass - and enable the adding and dropping of wavelengths at the node. Such add/drops may be rerouted traffic or provisioned new services.
As network traffic continues to grow, so do the degrees of a ROADM and the ports of its sub-systems. The degree of a ROADM is defined to the number of connections or fibre pairs it can support. In the diagram, a ROADM of degree three is shown.
A multicast switch-based 3-degree CDC ROADM. Source Lumentum.
It is rare to encounter more than five or six fibre routes leaving any given mesh node in a network, says Lumentum. “But in those fibre routes there is typically a large number of fibres - 64 or 128,” says Collings. “Operators deploy a conduit of fibre between cities.”
When the C-band fills up, an operator will light another fibre pair, taking up another of the ROADM’s degrees. ROADMs built today have 16 degrees. And since a fibre’s C-band can occupy some 30 terabits of data, this is how 400-terabit mesh nodes will be achieved.
“That is a pretty big node but that is the end [of life] capacity,” says Collings. “I don’t think you will find a 400-terabit node today but we build our networks so that they get there, five to eight years from when they are deployed.”
This raises another issue: the length of time it takes for any generational change of a ROADM design to take hold in the network.
“When a new approach comes along, it takes a couple of years for everyone to figure out how they will use it,” says Collings. Then, once a decision is made, it takes another two years to deploy followed by five to eight years before the ROADM node is filled.
“Nothing happens quickly in this business,” says Collings. “But the upside, from a business point of view, is that as things are designed in, they have a long deployment cycle.”
Lumentum illustrates the point with its own products.
The company is seeing growing demand for its dual TrueFlex WSS deployed in route-and-select ROADM architectures. “But we are still seeing growth on the older broadcast-and-select architectures underpinned by singe 1x9 WSSes,” says James Goodchild, director, product line management for wavelength management products at Lumentum.
CDC ROADMs
A colourless, directionless and contentionless (CDC) ROADM uses a twin multicast switch for the wavelength add and drop functions. The input fibre to each degree’s WSS is connected to the output path WSS of each of the ROADM’s other degrees. The input WSS also connects to the drop multicast switch (see diagram above).
Using a WSS on the input path means that only wavelengths of interest are routed to the WSS’ output ports. Hence the ROADM’s reference as a route-and-select architecture.
Using a 1xN splitter array instead of a WSS for the input path results in a broadcast-and-select ROADM. Here, the input fibre’s wavelengths are broadcast to all the N output ports. The high optical loss associated with the splitters is the main reason why CDC ROADM designs have transitioned to the WSS-based route-and-select architecture.
This second-generation switch is what will take us into the 100 gigabaud and super-channel era of network scalability
However, there is still an optical loss issue to be contended with, introduced by the add or drop multicast switch. Accordingly, along with the twin multicast switch are two arrays of erbium-doped fibre amplifiers (EDFAs). One EDFA array is on the drop ports to the MxN multicast switch and the second amplifier array boosts the outputs of the add-path multicast switch before their transmission into the network.
The MxN multicast switch comprises 1xN splitter arrays, N being the number of add-drop ports, and Mx1 selection switches where M is the number of directions the ROADM supports. A typical multicast switch is 8x16: eight being the ROADM’s number of directions and 16 the drop-port count.
Each of the N splitter arrays sends the signals on a drop port to all the Mx1 selection switches where each one pulls off the channel to be dropped. Having a selection switch at each of the multicast switch’s N drop ports is what enables contentionless operation, the avoidance of a collision when the same wavelength is droppedat a node from different degree directions.
MxN switch
Lumentum’s decision to develop the MxN switch to replace the multicast switch follows its study to understand how optical transmission networks will evolve with continual traffic growth.
One development is the adoption of higher-baud-rate, higher-capacity coherent transmissions that require wider channel widths. A 400-gigabit wavelength requires a 75GHz channel compared to the standard 50GHz fixed grid used for 100- and 200-gigabit transmissions. Future transmission speeds of 800 gigabits will use two such channels or 150GHz of spectrum, while a 1 terabit signal is expected to occupy 300GHz of fibre spectrum. “This is how we anticipate coherent transmission evolving,” says Collings.
Moving to wider channels also benefits the ROADM’s cost. If operators continued to use 50GHz channels, the channel count would grow exponentially with the growth in traffic. In contrast, adopting wider channels means the add-drop port count grows only linearly with traffic. “Using wider channels, the advantage is you don't have to support 600 ports of add-drop in your ROADM networks,” says Collings.
But wider channels means greater amplification demands on the EDFA arrays, an issue that will only worsen over time.
Multicast switch-based designs don’t support the wider channels we know are coming
Losing the amp
Because the power spectral density is constant, the power in a channel increases proportionally with its width. For example, a 75GHz channel has 2dB more power compared to a 50GHz channel spacing, a 150GHz channel 5dB more while a 300GHz channel has an extra 8dB.
The EDFA array is engineered to handle the worst case power requirement that occurs when all 16 optical transceivers into the multicast switch go to the same ROADM degree. Here the EDFA must be able to boost all 16 channels.
For a multicast switch with 16 ports, 22dBm amplification is needed for a 150GHz channel which requires going from an uncooled pump design to a cooled pump one. Equally, 25dBm amplification is needed for 300GHz channels. And as the number of degrees grows, so do the demands on the amplification until no practical amplifier design is possible (see diagram).
The EDFA requirements to compensate for the optical loss of the multicast switch. The complexity of the EDFA design grows with the multicast switch's port count until it becomes insupportable. Source: Lumentum.
“This is not an issue today because we use very modest-sized channels and we engineer our systems to accommodate them,” says Collings. “But if you look forward, you realise they [multicast switch-based designs] don’t support the wider channels we know are coming.”
Using a WSS-based MxN switch solves this issue because, as with the input port WSS of a route-and-select architecture, the switch has a lower optical loss - under 8dB - compared to the 17dB of the splitter-based multicast switch.
The sub-8dB loss is below the threshold where amplification is needed: the optical signal is sufficiently strong at the drop port to be received, as are the added signals for transmission into the network. The resulting removal of the EDFAs simplifies greatly the complexity, size and cost of the CDC ROADM.
“The MxN is a WSS - it’s a router - so it sends all of the light in the direction it is supposed to go,” says Collings. “You can push through the MxN switch channels of any width and of any power because there is no amplifier that needs to be there and be designed appropriately."
The resulting second-generation CDC ROADM design is shown below.
Source: Lumentum
Lumentum's Goodchild says the 8x24 twin implementation of the MxN switch will be available in the first quarter of 2019.
“Certain systems vendors already have access to samples,” says Goodchild.
Further reading
2D WSSes, click here
ROADMs and their evolving amplification needs, click here
Adding an extra dimension to ROADM designs
U.K. start-up ROADMap Systems, a developer of wavelength-selective switch technology, has completed a second round of funding. The amount is undisclosed but the start-up is believed to have raised several million dollars to date.
Karl HeeksThe company will use the funding to develop a prototype of its two-dimensional (2D) optical beam-steering technique to integrate 24 wavelength-selective switches (WSSes) within a single platform.
The WSS is a key building block used within reconfigurable optical add-drop multiplexers (ROADMs).
The company’s WSS technology uses liquid crystal on silicon (LCOS) technology, the basis of existing WSS designs from the likes of Finisar and Lumentum. However, the start-up has developed a way to steer beams in 2D whereas current WSSes operate in a single dimension only.
The Cambridge-based company’s pre-production prototype will integrate 24,1x12 WSSes within a single package. The platform promises service providers ROADM designs that deliver space, power consumption and operational cost savings as well as systems advantages.
Wavelength-selective switch
A WSS takes wavelength-division multiplexed (WDM) channels from an input fibre and distributes them as required across an array of output fibres. Typical WSS configurations include a 1x9 - a one input fibre port and nine output ports - and a 1x20.
Current WSS designs comprise a diffraction grating, a cylindrical lens and an LCOS panel that is used to deflect the light channels to the required output fibres.
The diffraction grating separates the WDM channels while the cylindrical lens produces an elongated projection of the channels onto the LCOS device. The panel’s liquid crystals are oriented in such a way to direct the projected light channels to the appropriate output fibres. The orientation of the arrays of liquid crystals that perform the various steerings are holograms.
Commercial WSSes use the LCOS panel to steer in one dimension only: left or right. This means the output fibres are arranged in an array and the number of fibres is limited by the total deflection the LCOS can achieve. ROADMap Systems has developed a technique that produces holograms on the LCOS panel that steer light in two dimensions: left and right, up and down and diagonally.
Moreover, the holograms are confined to a small area of the panel, far fewer pixels than the elongated beams of a 1D WSS. Such confinement allows multiple light beams to be steered to the output fibre bundles.
“You use a much smaller area of the LCOS to bend things in 2D,” says Karl Heeks, CEO at ROADMap Systems.
Platform demonstrator
ROADMap System’s key intellectual property is its know-how to create the steering pattern - the hologram - programmed onto the LCOS panel.
The 2D WSS system requires calibration to create the precision holograms. The calibration data is generated during the device’s manufacture and forms the input to an algorithm that creates the holograms needed for the LCOS to steer accurately the traffic to the output fibres.
You use a much smaller area of the LCOS to bend things in 2D
ROADMap Systems has demonstrated its 2D steering technology to service providers, system vendors and optical subsystem players.
Now, the company is working to build the 24, 1x12 WSSs on an optical bench which it expects to complete by the year-end. The start-up is also creating the calibration software used for 2D beam steering as well as a user interface to allow networking staff to set up their required connections.
The first pre-production packaged systems – each one comprising a 4K LCOS panel and 312 fibres - are expected for delivery for trialling in 2019. The start-up is reluctant to give a firm date as it is still exploring design options. For example, ROADMap Systems has an improved lower-loss, more compact fibre coupling design but it has yet to decide whether to incorporate it or its existing design for its platform.
“We are not intending the prototype to go into a system within the network,” says Heeks. “It is more a vehicle to illustrate its capabilities.”
System benefits
The main benefit of ROADMap Systems’ 2D beam-steering WSS architecture is not so much its optical performance; the start-up expects its design to match the optical performance of existing 1D WSSes. Rather, there are architectural benefits besides the obvious integration and cost benefits of putting 24 WSSes in one platform.
The first system advantage is the ability to use the many WSSes to implement ROADMs of several degrees including the ROADM’s add-drop architecture. A two-degree ROADM handles east and west fibre pairs while a three-degree ROADM adds north-facing traffic as well.
A ROADM architecture using 1xN splitters as part of the multicast switch. Source: ROADMap Systems.
To add and drop light-paths, a multicast switch is used (shown in green in the diagram above). The multicast switch can be implemented using optical splitters, however, due to their loss, optical amplifiers are needed to boost the signals, adding to the overall cost and system complexity.
WSSes can be used instead of the splitters as part of the multicast switch architecture such that optical amplification is not needed; the optical loss the WSS stage adds being much lower than the splitters. Removing optical amplification impacts significantly the overall ROADM cost (see diagram below).
A ROADM architecture using 1xN WSSes as part of the multicast switch. Source: ROADMap Systems.
The integrated platform’s large number of WSSes will ease the implementation of the latest generation of ROADMs that are colourless, directionless and contentionless.
A colourless ROADM decouples the wavelength-dependency such that a light-path can be used on any of the network interface ports. Directionless refers to having full flexibility in the routeing of a light-path to any of the ports. Lastly, contentionless means non-blocking, where the same wavelength channel can be accommodated across all the degrees of the ROADM without contention.
And being LCOS-based, ROADMap’s WSSes also support a flexible grid enabling the ROADM to support channels such as coherent transmissions above 200 gigabit-per-second that do not conform to the rigid 50GHz-wide ITU grid spacings.
The second system advantage of the platform is that with its many WSSes, it can route and add-drop wavelengths across both the C and L-bands. However, the company is not planning to implement this feature in its preproduction prototype.
Next steps
ROADMap Systems says it is focussed on producing and testing its pre-production prototype. A further round of investment will be needed to turn the design into a commercial product.
“We believe that such a highly-integrated architecture will offer immediate performance and economic benefits to many teleccom applications,” says Heeks. “It is also well positioned for datacentre – DCI - applications where data needs to be routed between distributed datacentres linked by parallel fibres."
OFC 2015 digest: Part 1
- Several vendors announced CFP2 analogue coherent optics
- 5x7-inch coherent MSAs: from 40 Gig submarine and ultra-long haul to 400 Gig metro
- Dual micro-ITLAs, dual modulators and dual ICRs as vendors prepare for 400 Gig
- WDM-PON demonstration from ADVA Optical Networking and Oclaro
- More compact and modular ROADM building blocks
JDSU also showed a dual-carrier coherent lithium niobate modulator capable of 400 Gig for long-reach applications. The company is also sampling a dual 100 Gig micro-ICR also for multiple sub-channel applications.
Avago announced a micro-ITLA device using its external cavity laser that has a line-width less than 100kHz. The micro-ITLA is suited for 100 Gig PM-QPSK and 200 Gig 16-QAM modulation formats and supports a flex-grid or gridless architecture.
ROADMs and their evolving amplification needs
Technology briefing: ROADMs and amplifiers
Oclaro announced an add/drop routing platform at the recent OFC/NFOEC show. The company explains how the platform is driving new arrayed amplifier and pumping requirements.
A ROADM comprising amplification, line-interfaces, add/ drop routing and transponders. Source: Oclaro
Agile optical networking is at least a decade-old aspiration of the telcos. Such networks promise operational flexibility and must be scalable to accommodate the relentless annual growth in network traffic. Now, technologies such as coherent optical transmission and reconfigurable optical add/drop multiplexers (ROADMs) have reached a maturity to enable the agile, mesh vision.
Coherent optical transmission at 100 Gigabit-per-second (Gbps) has become the base currency for long-haul networks and is moving to the metro. Meanwhile, ROADMs now have such attributes as colourless, directionless and contentionless (CDC). ROADMs are also being future-proofed to support flexible grid, where wavelengths of varying bandwidths are placed across the fibre's spectrum without adhering to a rigid grid.
Colourless and directionless refer to the ROADM's ability to transmit or drop any light path from any direction or degree at any network interface port. Contentionless adds further flexibility by supporting same-colour light paths at an add or a drop.
"You can't add and drop in existing architectures the same colour [light paths at the same wavelength] in different directions, or add the same colour from a given transponder bank," says Bimal Nayar, director, product marketing at Oclaro's optical network solutions business unit. "This is prompting interest in contentionless functionality."
The challenge for optical component makers is to develop cost-effective coherent and CDC-flexgrid ROADM technologies for agile networks. Operators want a core infrastructure with components and functionality that provide an upgrade path beyond 100 Gigabit coherent yet are sufficiently compact and low-power to minimise their operational expenditure.
ROADM architectures
ROADMs sit at the nodes of a mesh network. Four-degree nodes - the node's degree defined as the number of connections or fibre pairs it supports - are common while eight-degree is considered large.
The ROADM passes through light paths destined for other nodes - known as optical bypass - as well as adds or drops wavelengths at the node. Such add/drops can be rerouted traffic or provisioned new services.
Several components make up a ROADM: amplification, line-interfaces, add/drop routing and transponders (see diagram, above).
"With the move to high bit-rate systems, there is a need for low-noise amplification," says Nayar. "This is driving interest in Raman and Raman-EDFA (Erbium-doped fibre amplifier) hybrid amplification."
The line interface cards are used for incoming and outgoing signals in the different directions. Two architectures can be used: broadcast-and-select and route-and select.
With broadcast-and-select, incoming channels are routed in the various directions using a passive splitter that in effect makes copies the incoming signal. To route signals in the outgoing direction, a 1xN wavelength-selective switch (WSS) is used. "This configuration works best for low node-degree applications, when you have fewer connections, because the splitter losses are manageable," says Nayar.
For higher-degree node applications, the optical loss using splitters is a barrier. As a result, a WSS is also used for the incoming signals, resulting in the route-and-select architecture.
Signals from the line interface cards connect to the routing platform for the add/drop operations. "Because you have signals from any direction, you need not a 1xN WSS but an LxM one," says Nayar. "But these are complex to design because you need more than one switching plane." Such large LxM WSSes are in development but remain at the R&D stage.
Instead, a multicast switch can be used. These typically are sized 8x12 or 8x16 and are constructed using splitters and switches, either spliced or planar lightwave circuit (PLC) based .
"Because the multicast switch is using splitters, it has high loss," says Nayar. "That loss drives the need for amplification."
Add/drop platform
With an 8-degree-node CDC ROADM design, signals enter and exit from eight different directions. Some of these signals pass through the ROADM in transit to other nodes while others have channels added or dropped.
In the Oclaro design, an 8x16 multicast switch is used. "Using this [multicast switch] approach you are sharing the transponder bank [between the directions]," says Nayar.
The 8-degree node showing the add/drop with two 8x16 multicast switches and the 16-transponder bank. Source: Oclaro
A particular channel is dropped at one of the switch's eight input ports and is amplified before being broadcast to all 16, 1x8 switches interfaced to the 16 transponders.
It is the 16, 1x8 switches that enable contentionless operation where the same 'coloured' channel is dropped to more than one coherent transponder. "In a traditional architecture there would only be one 'red' channel for example dropped as otherwise there would be [wavelength] contention," says Nayar.
The issue, says Oclaro, is that as more and more directions are supported, greater amplification is needed. "This is a concern for some, as amplifiers are associated with extra cost," says Nayar.
The amplifiers for the add/drop thus need to be compact and ideally uncooled. By not needing a thermo-electrical cooler, for example, the design is cheaper and consumes less power.
The design also needs to be future-proofed. The 8x16 add/ drop architecture supports 16 channels. If a 50GHz grid is used, the amplifier needs to deliver the pump power for a 16x50GHz or 800GHz bandwidth. But the adoption of flexible grid and super-channels, the channel bandwidths will be wider. "The amplifier pumps should be scalable," says Nayar. "As you move to super-channels, you want pumps that are able to deliver the pump power you need to amplify, say, 16 super-channels."
This has resulted in an industry debate among vendors as to the best amplifier pumping scheme for add/drop designs that support CDC and flexible grid.
EDFA pump approaches
Two schemes are being considered. One option is to use one high-power pump coupled to variable pump splitters that provides the required pumping to all the amplifiers. The other proposal is to use discrete, multiple pumps with a pump used for each EDFA.
Source: Oclaro
In the first arrangement, the high-powered pump is followed by a variable ratio pump splitter module. The need to set different power levels at each amplifier is due to the different possible drop scenarios; one drop port may include all the channels that are fed to the 16 transponders, or each of the eight amplifiers may have two only. In the first case, all the pump power needs to go to the one amplifier; in the second the power is divided equally across all eight.
Oclaro says that while the high-power pump/ pump-splitter architecture looks more elegant, it has drawbacks. One is the pump splitter introduces an insertion loss of 2-3dB, resulting in the pump having to have twice the power solely to overcome the insertion loss.
The pump splitter is also controlled using a complex algorithm to set the required individual amp power levels. The splitter, being PLC-based, has a relatively slow switching time - some 1 millisecond. Yet transients that need to be suppressed can have durations of around 50 to 100 microseconds. This requires the addition of fast variable optical attenuators (VOAs) to the design that introduce their own insertion losses.
"This means that you need pumps in excess of 500mW, maybe even 750mW," says Nayar. "And these high-power pumps need to be temperature controlled." The PLC switches of the pump splitter are also temperature controlled.
The individual pump-per-amp approach, in contrast, in the form of arrayed amplifiers, is more appealing to implement and is the approach Oclaro is pursuing. These can be eight discrete pumps or four uncooled dual-chip pumps, for the 8-degree 8x16 multicast add/drop example, with each power level individually controlled.
Source: Oclaro
Oclaro says that the economics favour the pump-per-amp architecture. Pumps are coming down in price due to the dramatic price erosion associated with growing volumes. In contrast, the pump split module is a specialist, lower volume device.
"We have been looking at the cost, the reliability and the form factor and have come to the conclusion that a discrete pumping solution is the better approach," says Nayar. "We have looked at some line card examples and we find that we can do, depending on a customer’s requirements, an amplified multicast switch that could be in a single slot."
A FOX-C approach to flexible optical switching
Flexible switching of high-capacity traffic carried over ’super-channel' dense-wavelength division multiplexing wavelengths is the goal of the European Commission Seventh Framework Programme (FP7) research project.
The €3.6M FOX-C (Flexible optical cross-connect nodes) will develop a flexible spectrum reconfigurable optical add/drop multiplexer (ROADM) for 400Gbps and one Terabit optical transmission. The ROADM will be designed not only to switch super-channels but also the carrier constituent components.
Companies involved in the project include operator France Telecom and optical component player Finisar. However, no major European system vendor is taking part in the FOX-C project although W-Onesys, a small system vendor from Spain, is participating.
“We want to transfer to the optical layer the switching capability”
Erwan Pincemin, FT-Orange
“It is becoming more difficult to increase the spectral efficiency of such networks,” says Erwan Pincemin, senior expert in fibre optic transmission at France Telecom-Orange. “We want to increase the advantages of the network by adding flexibility in the management of the wavelengths in order to adapt the network as services evolve.”
FOX-C will increase the data rate carried by each wavelength to achieve a moderate increase in spectral efficiency. Pincemin says such modulation schemes as orthogonal frequency division multiplexing (OFDM) and Nyquist WDM will be explored. But the main goal is to develop flexible switching based on an energy efficient and cost effective ROADM design.
The ROADM’s filtering will be able to add and drop 10 and 100 Gigabit sub-channels or 400 Gigabit and 1 Terabit super-channels. By using the developed filter to switch optically at speeds as low as 10 Gigabit, the aim is to avoid having to do the switching electrically with its associated cost and power consumption overhead. “We want to transfer to the optical layer the switching capability,” says Pincemin.
While the ROADM design is part of the project’s goals, what is already envisaged is a two-stage pass-through-and-select architecture. The first stage, for coarse switching, will process the super-channels and will be followed by finer filtering to extract (drop) and insert (add) individual lower-rate tributaries.
The project started in Oct 2012 and will span three years. The resulting system testing will take place at France-Telecom Orange's Lab in Lannion, France.
Project players
The project’s technical leader is the Athens Institute of Technology (AIT), headed by Prof. Ioannis Tomkos, while the administrator leader is the Greek company Optronics Technologies.
Finisar will provide the two-stage optical switch while France Telecom-Orange will test the resulting ROADM and will build the multi-band OFDM transmitter and receiver to evaluate the design.
Athens Institute of Technology will work with Finisar on the technical aspects and in particular a flexible networking architecture study. The Hebrew University is working with Finisar on the design and the building of the ultra-selective adaptive optical filter, and has expertise is free-space optical systems. The Spanish firm, W-Onesys, is a system integration specialist and will also work with Finisar to integrate its wavelength-selective switch for the ROADM. Other project players include Aston University, Tyndall National Institute and the Karlsruhe Institute of Technology.
No major European system vendor is taking part in the FOX-C project. According to Pincemin this is regrettable although he points out that the equipment players are involved in other EC FP7 projects addressing flexible networking.
He believes that their priorities are elsewhere and that the FOX-C project may be deemed as too forward looking and risky. “They want to have a clear return on investment on their research,” says Pincemin.
JDSU's Brandon Collings on silicon photonics, optical transport & the tunable SFP+
JDSU's CTO for communications and commercial optical products, Brandon Collings, discusses reconfigurable optical add/drop multiplexers (ROADMs), 100 Gigabit, silicon photonics, and the status of JDSU's tunable SFP+.
"We have been continually monitoring to find ways to use the technology [silicon photonics] for telecom but we are not really seeing that happen”
Brandon Collings, JDSU
Brandon Collings highlights two developments that summarise the state of the optical transport industry.
The industry is now aligned on the next-generation ROADM architecture of choice, while experiencing a ’heavy component ramp’ in high-speed optical components to meet demand for 100 Gigabit optical transmission.
The industry has converged on the twin wavelength-selective switch (WSS) route-and-select ROADM architecture for optical transport. "This is in large networks and looking forward, even in smaller sized networks," says Collings.
In a route-and-select architecture, a pair of WSSes reside at each degree of the ROADM. The second WSS is used in place of splitters and improves the overall optical performance by better suppressing possible interference paths.
JDSU showcased its TrueFlex portfolio of components and subsystems for next-generation ROADMs at the recent European Conference on Optical Communications (ECOC) show. The company first discussed the TrueFlex products a year ago. "We are now in the final process of completing those developments," says Collings.
Meanwhile, the 100 Gigabit-per-second (Gbps) component market is progressing well, says Collings. The issues that interest him include next-generation designs such as a pluggable 100Gbps transmission form factor.
Direct detection and coherent
JDSU remains uncertain about the market opportunities for 100Gbps direct-detection solutions for point-to-point and metro applications. "That area remains murky," says Collings. "It is clearly an easy way into 100 Gig - you don't have to have a huge ASIC developed - but its long-term prospects are unclear."
The price point of 100Gbps direct-detection, while attractive, is competing against coherent transmission solutions which Collings describes as volatile. "As coherent becomes comparable [in cost], the situation will change for the 4x25 Gig [direct detection] quite quickly," he says. "Coherent seems to be the long-term, robust cost-effective way to go, capturing most of the market."
At present, coherent solutions are for long-haul that require a large, power-consuming ASIC. Equally the accompanying optical components - the lasers and modulators - are also relatively large. For the coherent metro market, the optics must become cheaper and smaller as must the coherent ASIC.
"If you are looking to put that [coherent ASIC and optics] into a CFP or CFP2, the problem is based on power; cost is important but power is the black-and-white issue," says Collings. Engineers are investigating what features can be removed from the long-haul solution to achieve the target 15-20W power consumption. "That is pretty challenging from an ASIC perspective and leaves little-to-no headroom in a pluggable," says Collings.
The same applies to the optics. "Is there a lesser set of photonics that can sit on a board that is much lower cost and perhaps has some weaker performance versus today's high-performance long-haul?" says Collings. These are the issues designers are grappling with.
Silicon photonics
Another area in flux is the silicon photonics marketplace. "It is a very fluid and active area," says Collings. "We are not highly active in the area but we are very active with outside organisations to keep track of its progress, its capabilities and its overall evolution in terms of what the technology is capable of."
The silicon photonics industry has shifted towards datacom and interconnect technology in the last year, says Collings. The performance levels silicon photonics achieves are better suited to datacom than telecom's more demanding requirements. "We have been continually monitoring to find ways to use the technology for telecom but we are not really seeing that happen,” says Collings.
Tunable SFP+
JDSU demonstrated its tunable laser in an SFP+ pluggable optical module at the ECOC exhibition.
The company was first to market with the tunable XFP, claiming it secured JDSU an almost two-year lead in the marketplace. "We are aiming to repeat that with the SFP+," says Collings.
The SFP+ doubles a line card's interface density compared to the XFP module. The SFP+ supports both 10Gbps client-side and wavelength-division multiplexing (WDM) interfaces. "Most of the cards have transitioned from supporting the XFP to the SFP+," says Collings. This [having a tunable SFP+] completes that portfolio of capability."
JDSU has provided samples of the tunable pluggable to customers. "We are working with a handful of leading customers and they typically have a preference on chirp or no-chirp [lasers], APD [avalanche photo-diode] or no APD, that sort of thing," says Collings.
JDSU has not said when it will start production of the tunable SFP+. "It won't be long," says Collings, who points out that JDSU has been demonstrating the pluggable for over six months.
The company plans a two-stage rollout. JDSU will launch a slightly higher power-dissipating tunable SFP+ "a handful of months" before the standard-complaint device. "The SFP+ standard calls for 1.5W but for some customers that want to hit the market earlier, we can discuss other options," says Collings.
Further reading
OFC/NFOEC 2012: Some of the exhibition highlights
A round-up of some of the main announcements and demonstrations at the recent OFC/NFOEC 2012 exhibition and conference.
100 Gigabit coherent
Finisar demonstrated its first 100 Gigabit coherent receiver transponder. The 5x7inch dual-polarisation, quadrature phase-shift keying (DP-QPSK) module complies with the Optical Internetworking Forum's (OIF) multi-source specification. The companies joins Fujitsu Optical Components, Opnext and Oclaro that have already detailed their 100 Gigabit coherent modules. Since OFC/NFOEC, Oclaro and Opnext have announced their intention to merge.
"We can take off-the-shelf DSP technology and match it with vertically-integrated optics and come up with a module that is cost effective while enabling higher density for system vendors," says Rafik Ward, vice president of marketing at Finisar. "This will start the shift away from the system vendors' proprietary line cards."
Opnext announced it has demonstrated interoperability between its OMT-100 100 Gigabit-per-second (Gbps) coherent module and 100 Gigabit systems from Fujitsu Optical Systems and NEC. All three designs use NTT Electronics' (NEL) DSP-ASIC coherent receiver chip. "For those that use the same NEL modem chip, we can interoperate with each other," says Ross Saunders, general manager, next-generation transport for Opnext Subsystems.
They come back to folks like us and say: 'If you can hit this price point, then we will use you'
Ross Saunders, Opnext
Oclaro's MI 8000XM 100Gbps module also uses the NEL DSP-ASIC but was not part of the interoperability test sponsored by Japanese operator, NTT.
Oclaro announced its 100Gbps coherent module is now being manufactured using all its own optical components. These include a micro integrated tunable laser assembly (ITLA) - the latest ratified MSA that is more compact and has a higher output power, its modulator and its coherent receiver module.
Using its components enables the company to control performance-cost tradeoffs, says Per Hansen, vice president of product marketing, optical networks solutions at Oclaro: "This [vertical integration] gives us a flexibility we didn’t have in the past."
Finisar is not saying which merchant DSP-ASIC it is using. But like the NEL device, the DSP-ASIC supports soft-decision forward error correction (SD-FEC) to achieve a reach of over 2,000km.
Meanwhile, the module makers' 100Gbps modules are starting to be shipped to customers.
"We shipped [samples] to four customers last quarter and we are probably going to ship to another four or five by the end of this quarter," says Opnext's Saunders.
Opnext says nearly all of its early customers do not have their own in-house 100Gbps developments. However, the systems vendors that have internal 100Gbps programmes have designed their line cards using the same 168-pin interface. This allows them to replace their own 100Gbps daughter cards with a merchant 5x7-inch module.
"This [vertical integration] gives us a flexibility we didn’t have in the past."
Per Hansen, Oclaro
The company also announced its OTS-100FLX 100Gbps muxponder, transponder and regenerator line cards that use the OTM-100 module and which slot into its OTS-4000 chassis. The chassis supports eight 100Gbps cards. Opnext's smaller 4RU OTS-mini platform hosts two 100Gbps line cards, mounted horizontally. Over half of Opnext's revenues are from subsystems sales which it brands and sells to system vendors.
As for the other 100Gbps transponder makers, Oclaro is sending out its first module samples now. Finisar says its module will be generally available by the year-end, while Fujitsu Optical Components' module was released in April.
Optical components for 200Gbps DP-QPSK
u2t Photonics announced its latest 64Gbaud photo-detector that points to the next speed shift in line-side transmission. The photo-detector is one key building block to the eventual development of a single-carrier DP-QPSK capable of 200Gbps or using 16-QAM, 400Gbps.
"We can already support the higher interface speed and data throughput"
Jens Fiedler, u2t Photonics
"System companies are looking for two things: to increase the baud rate and to use more complex modulation schemes," says Jens Fiedler, vice president sales and marketing at u2t Photonics. "[With this announcement] from the optical component perspective, we can already support the higher interface speed and data throughput."
1x23 Wavelength-selective switch
Oclaro announced a 1x23 wavelength selective switch at OFC. According to Oclaro, the 1x23 WSS has come about due to the operators' desire to support 12-degree nodes: an input port (1 degree) and through-connections on 11 other ports. The remaining 12 [of the WSS's 24 ports] are used as drop ports.
"If for each of those ports you have a fan-out that is steerable to 8 ports, you have 12x8 or 96 as the total channels you can support for a full add-drop," says Hansen. Such a 12-degree, 96-channel requirement was set by operators early on, or at least it was an industry desire, says Hansen.
Switching elements that address these drop requirements - multicast switches - were announced by NeoPhotonics and Enablence Technologies at OFC. The switches, planar lightwave circuit (PLC) hybrid integration designs, implement 8x16 multicast switches.
"The multicast switch takes signals from eight different inputs - 8 different directions in a ROADMs node and distributes those signals to up to 16 drop ports," says Ferris Lipscomb, vice president of marketing at NeoPhotonics.
Such PLC designs are complex, comprising power splitters, waveguide switching, variable optical attenuators and photo-detectors for channel monitoring.
According to NeoPhotonics, the number of optical functions used to implement the multicast switch is in the hundreds.
Enablence already has 8x8 and 8x12 multicast switches and has launched its 8x16 device. Although the company is a hybrid PIC specialist and has PLC technology, it uses polymer PLCs for the multicast designs, claiming they are lower power. NEL is another company offering 8x8 and 8x12 multicast switches.
Passive optical networking
Finisar also demonstrated a mini-PON network, highlighting its optical line terminal (OLT) transceivers, splitters and its latest GPON-stick, an GPON optical network unit built into an SFP. The demo involved using the ONU SFP transceiver in an Ethernet switch port as part of a PON network to deliver high-definition video and audio from the OLT to a high-definition TV.
The company also introduced two splitter products a 1:128 port splitter and a 2:64 (used for redundancy). These high-split ratios are being prepared for the advent of 10 Gigabit PON.
Enablence also demonstrated a WDM-PON 32-channel receiver module at OFC. "It takes 32 TO-can receivers and replaces them with a small module which includes the AWG (arrayed waveguide grating demultiplexer) and the 32 receivers," says Matt Pearson, vice president, technology, optical components division at Enablence Technologies. The design promises to increase system density by fitting two such receivers on a single blade.
Optical engines
Silicon photonics firm, Kotura, detailed its 100Gbps optical engine chip, implemented as a 4x25Gbps design. The optical engine consumes 5W and has a reach of at least 10km, making it suitable for requirements in the data centre including the 100 Gigabit Ethernet IEEE 100GBASE-LR4 standard.
"The 100Gbps chip - 5mmx6mm - is small enough to fit in the QSFP+ and emerging CFP4 optical modules
Arlon Martin, Kotura
Kotura demonstrated to select customers its optical engine. "We are not announcing the product yet," says Arlon Martin, vice president of marketing at Kotura.
Optical engines are used in several applications: pluggable modules on a system's face-plate, the optics at each end of an active optical cable, and for board-mounted embedded applications.
For embedded applications, the optical engine is mounted deeper within the line card, close to high-speed chips, for example, with the signals routed over fibre to the face-plate connector. Using optics rather than high-speed copper traces simplifies the printed circuit board design.Embedded optical engines will also be used for optical backplane-based platforms.
Kotura's silicon photonics-based optical engine integrates all the functions needed for the transmitter and receiver on-chip. These include the 25Gbps optical modulators and drivers, the 4:1 multiplexer and 1:4 demultiplexer and four photo-detectors. To create the lasers, an array of four gain blocks are coupled to the chip. Each of laser's wavelength, around 1550nm, is set using on-chip gratings.
The 100Gbps chip, measuring about 5mmx6mm, is small enough to fit in the QSFP+ and emerging CFP4 optical modules, says Martin. The QSFP+ is likely to be the first application for Kotura's 100Gbps optical engine, used to connect switches within the data centre.
Finisar demonstrated its own VCSEL-based board mount optical assembly - also an optical engine - to highlight the use of the technology for future optical backplanes.
The demonstration, involving Vario-optics and Huber + Suhner, included boards in a chassis. The board includes the optical engine coupled to polymer waveguides from Vario-optics which connect it to a backplane connector, built by Huber + Suhner. "The idea is to show what an integrated optical chassis will look like," says Ward.
Finisar's optical backplane demo using board-mounted optics. Source: Finisar
The optical engine comprises 24 channels - 12 transmitters at 10Gbps and 12 receivers in a single board-mounted package. The optics can operate at 10, 12, 14, 25 and 28Gbps, says Finisar. The connector allows the optical engines on different cards to interface via the waveguides. The advantage of polymer waveguides is that they are relatively easy to etch on printed circuit boards and since they replace fibre, they remove fibre management issues. However the technology needs to be proven before system vendors will use such waveguides as standard in their platforms.
Interconnect specialist Reflex Photonics demonstrated an 8.6Tbps optical backplane at OFC. The demonstrator uses Reflex's LightABLE optical engines to implement 864 point-to-point optical fibre links to achieve 8.6Tbps in a single chassis.
The optical fabric comprises six layers of 12x12 fully connected broadcast meshes. Each line card supports 720Gbps into the optical backplane and 60Gbps direct bandwidth between any two cards.
32G Fibre channel
Finisar also highlighted its 28Gbps VCSEL that will be used for the 32 Gigabit Fibre Channel standard. The actual line rate for 32Gbps Fibre Channel is 28.05Gbps. The VCSEL is packaged into a transmitter optical sub-assembly (TOSA) that fits inside a SFP+ module.
"We view 28Gbps VCSEL as strategic due to all the applications it will enable," says Ward.
Besides 32Gbps Fibre Channel, the high-speed VCSEL is suited for the next Infiniband data rate - enhanced data rate (EDR) at 4x25Gbps or 12x25Gbps. There is also standards work in the IEEE for a new 100Gbps Ethernet standard that can use 4x25Gbps VCSELs.
Further reading:
Gazettabyte's full OFC NFOEC 2012 coverage
LightCounting: Notes from OFC 2012: Onset of the Terabit Age
Ovum's OFC coverage