Capella: Why the ROADM market is a good place to be

Gazettabyte spoke with Larry Schwerin, CEO of Capella Intelligent Subsystems, about the ROADM market, the company's plans following its latest funding round, and the idea of a WSS-on-a-chip.

The reconfigurable optical add-drop multiplexer (ROADM) market has been the best performing segment of the optical networking market over the last year. According to Infonetics Research, ROADM-based wavelength division multiplexing (WDM) equipment grew 20% from Q2, 2010 to Q1, 2011 whereas the overall optical networking market grew 7%.

“It’s the Moore’s Law: Every two years we are doubling the capacity in terms of channel count and port count”

Larry Schwerin, Capella

The ROADM market has since slowed down but Larry Schwerin, CEO of wavelength-selective-switch (WSS) provider, Capella Intelligent Subsystems, says the market prospects for ROADMs remain solid.

Capella makes WSS products that steer and monitor light at network nodes, while the company’s core intellectual property is closed-loop control.  Its WSS products are compact, athermal designs based on MEMS technology that switch and monitor light.

Schwerin compares Capella to a plumbing company: “We clean out pipes and those pipes happen to be fibre-optics ones.”  The reason such pipes need ‘cleaning’ – to be made more efficient - is because of the content they carry. “It is bandwidth demand and the nature of the bandwidth which has changed dramatically, that is the fundamental driver here,” says Schwerin.

Increasingly the content is high-bandwidth video and streamed to end-user devices no longer confined to the home, while the video requested is increasingly user-specific. Such changes in the nature of content are affecting the operators’ distribution networks.

“Using Verizon as an example, they are now pushing 50 wavelengths per fibre in the metro,” says Schwerin. Such broad lanes of traffic arrive at network congestion points where certain fibre is partially used while other fibre is heavily used. “What they [operators] need is a vehicle that allows them to dynamically and remotely reassign those wavelengths on-the-fly,” says Schwerin. “That is what the ROADM does.”

Capella attributes strong ROADM sales to a maturing of the technology coupled with a price reduction. The technology also brings valuable flexibility at the optical layer. “It [ROADM] extends the life of the existing infrastructure, avoiding the need for capital to put new fibre in - which is the last thing the operators want to do,” says Schwerin. 

$20M funding

Capella raised US $20M in April as part of its latest funding round. The funding is being used for capital expansion and R&D. “We are working on new engine technology, new patentable concepts,” says Schwerin. “We were at Verizon a few weeks ago doing a world-first demo which we will be putting out as a press release.” For now the company will say that the demonstration is research-oriented and will not be implemented within ROADM systems anytime soon.

“You have to be competitive in this market, that is the downfall of our sector. People getting 30 or 40% gross margins and calling that a win – that is not a win - that is why this sector is in trouble”

One investor in the latest funding round is SingTel Innov8, the investment arm of the operator SingTel. Schwerin says it has no specific venture with the operator but that SingTel will gain insight regarding switching technologies due to the investment. “We will sit down with them and talk about their plans for network evolution and what is technologically possible,” says Schwerin, who points out that many of the carriers have lost contact with technologies since they shed their own large, in-house R&D arms.

Cappella offers two 1x9 WSS products and by the end of this year will also offer a 1x20 product. “It’s the Moore’s Law: Every two years we are doubling the capacity in terms of channel count and port count,” says Schwerin.

“We have a reasonable share of design wins shipping in volume - we have thousands of switches deployed throughout the world,” says Schwerin. “We are not of the size of a JDSU or a Finisar but our objective within the next 18 months is to capture enough market share that you would see us as a main supplier of that ilk.”

The CEO stresses that Capella’s presence a decade after the optical boom ended proves it is offering distinctive products. “Our whole business model is about innovation and differentiation,” says Schwerin.

But as a start-up how can Capella compete with a JDSU or a Finisar? “I have these conversations with the carriers: if all they are doing is looking for second or third sourcing of commodity product parts then there is no room for a company like a Capella.”  

The key is taking a dumb switch and turning it into a complete wavelength managed solution that can be easily added within the network. 

Schwerin also stresses the importance of ROADM specsmanship: wider lightpath channel passbands, lower insertion loss, smaller size, lower power consumption and competitive pricing: “You have to be competitive in this market, that is the downfall of our sector,” says Schwerin. “People getting 30 or 40% gross margins and calling that a win – that is not a win - that is why this sector is in trouble.”

Advanced ROADM features

There has been much discussion in the last year regarding the four advanced attributes being added to ROADM designs: colourless, directionless, contentionless and gridless or CDCG for short.

Interviewing six system vendors late last year, while all claimed they could support CDCG features, views varied as to what would be needed and by when. Meanwhile all the system vendors were being cautious until it was clearer as to what operators needed.

Schwerin says that what the operators really want is a ‘touchless’ ROADM.  Capella says its platform is capable of supporting each of the four attributes and that the company has plans for implementing each one. “Just because the carriers say they want it, that doesn’t mean that they are willing to pay for it,” says Schwerin. “And given the intense pricing pressure our system friends are in, they are rightly being cautious.”

Capella says that talking to the carriers doesn’t necessarily answer the issue since views vary as to what is needed. “The one [attribute] that seems clearest of all is colourless,” says Schwerin. And colourless is served using higher-port-count WSSs.

The directionless attribute is more a question of implementation and the good news is that it requires more WSSs, says Schwerin.  Contentionless addresses the issue of wavelength blocking and is the most vague, a requirement that has even “faded away a bit”.  As for gridless, that may be furthest out as it has ramifications in the network.

Schwerin says that Capella is seeing requests for reduced WSS switching times as well as wavelength tracking, tagging a wavelength whose signature can be identified optically and which is useful for network restoration and when wavelengths are passed between carriers’ networks.

Roadmap

In terms of product plans, Capella will launch a 1x20 WSS product later this year.  The next logical step in the development of WSS technology is moving to a solid-state-based design. 

“All of the the technologies out there today– liquid crystal, MEMS, liquid-crystal-on-silicon - are all free space [designs],” says Schwerin. “We have a solid-state engine in the middle [of our WSS] and we are down to five photonic-integrated-circuit components so the obvious next stage is silicon photonics.”

Does that mean a waveguide-based design? “Something of that form – it may not be a waveguide solution but something akin to that - but the idea is to get it down to a chip,” says Schwerin. “We are not pure silicon photonics but we are heading that way.”

Such a compact chip-based WSS design is probably five years out, concludes Schwerin.

Further information:

A Fujitsu ROADM discussion with Verizon and Capella – a Youtube 30-min video 

Comments

[Roy Rubenstein]

Nice article! My company, Optelian, is very focused on the ROADM market as well as other technology developments in the next-gen optical landscape.

Posted by Stephen Loudermilk on LinkedIn's Optical Networking Group.