OFC/NFOEC 2013 industry reflections - Part 3

Joe Berthold, vice president of network architecture, Ciena

The two topics that received the most attention, judging from session attendance and discussion in the hallways, were silicon photonics and software-defined networking (SDN). I predict that next year those who wish to capitalise on this popularity wave will be submitting papers on SDN-enabled silicon photonics.

More seriously, though, there remains vigorous debate about the relative importance of III-V integrated optics and silicon photonics, and I look forward to seeing how this evolves in the marketplace.

"Some of the SDN-related talks from the global research and education community were very good. They have been pioneers in making high capacity optical networks dynamic, and we have much to learn from them as they have several years experience building and operating SDNs, even before the term existed."

With respect to SDN and service providers, it is going to be several years before we see a true, SDN-enabled network as there are many other issues that need to be addressed.

This is one of the reasons Ciena is taking a lead role in the Open Networking Foundation's investigation of applying OpenFlow or the like at the optical layers. I thought some of the SDN-related talks from the global research and education community were very good. They have been pioneers in making high capacity optical networks dynamic, and we have much to learn from them as they have several years experience building and operating SDNs, even before the term existed.

"One of the most interesting commercial developments to watch in the coming years related to 100 Gig is the work that has begun on pluggable coherent analogue optical modules"

There was also quite a bit of buzz about 100 Gig deployments. It was nice to hear one of the industry analysts refer to 2013 as the year of 100 Gig as this is an area where Ciena has been quite successful.

I did not see or hear of any dramatic advances reported at the conference. What I did see, in talks and on the show floor, was a broad base of technology development that will lead to increased system density and lower cost and power.

On the client side, many companies showed 100 Gig CFP2 modules, and there was quite a bit of talk and demonstrations of technology building blocks that will lead to even smaller size.

Another optical networking topic that means many different things to different people was flexible grids and flexible transmission formats. From speaking with a number of network operators, it seems there is an appreciation for the future-proofing benefit of flexible grid ROADMs, but a recognition that the spectral efficiency gains to be had are quite limited, especially in a ROADM mesh network. So they are emerging as a nice-to-have feature but not a must-have-at-any-price feature.

Another 'flex' concept is flex-transceivers. The flavour of flex-transceivers that seem by most I spoke with to be practical are those that maintain a fixed baud rate but vary modulation format, say from BPSK to QPSK, 8PSK, 16QAM and perhaps beyond, to fit different distance applications.

I think one of the most interesting commercial developments to watch in the coming years related to 100 Gig is the work that has begun on pluggable coherent analogue optical modules, likely to emerge in a CFP2 form factor. I view this as a major next step the industry will take to reduce the cost and increase the density of coherent interfaces on switches and transmission systems.

The OIF did the industry a great service in pulling together a set of interoperable building blocks that form the photonic foundation of 100 Gig solutions today. The next step is to integrate these pieces and place them in a pluggable module. There is yet no formal project with this goal, but discussions are underway.

Watch this space...

Karen Liu, principal analyst components, Ovum.

There was a real sense of openness to new directions even as a lot of short-term activity continues to focus on getting 100 Gig to full maturity. Instead of pitching their favourite directions, some people actually solicited more ideas.

 "One trend to watch is the battle between VCSELs and silicon photonics"

Directions that seemed promising but unformed last year got a bit firmed up. Connections are being made from the application down to the device technology. What had been wacky ideas previously are being taken seriously:

• Optical circuit switching looks like it will have a place in conjunction with Ethernet switching.
• Spatial division multiplexing is the hot research topic. I like the work that Bell Labs is doing, particularly where the add/drop increment ties together multiple cores of the same wavelength so compensation algorithms can take advantage of similar environmental history.  This is moving past the physics, to thinking about network architecture.
• Monolithic integration of electronics with photonics. Early stages still and primarily around the drivers. But as this is motivated by power consumption, it seems like a solid direction that will have legs. 

One trend to watch is the battle between VCSELs (vertical-cavity surface-emitting lasers) and silicon photonics. Conventional wisdom was that VCSELS were for multi-mode and silicon photonics for single-mode but both have crossed over into the other's space.

Martin Guy, vice-president of product management and technology Teraxion 

There were several noteworthy developments. In particular, silicon photonics has started to show its promises as new products are introduced:

• Cisco announced its 100 Gig CPAK transceiver following the Lightwire acquisition
• Kotura showed its 100 Gig WDM QSFP package with only 3.5 W of power consumption.
• Luxtera demonstrated a 100 Gig QSFP package using four fibre pairs, each [fibre] carrying 25Gbps.
• Teraxion introduced its small form factor coherent receiver based on silicon photonics 

Silicon photonics was also largely discussed at the technical conference and very impressive results were demonstrated. Most notably, Cisco and Alcatel-Lucent presented results on silicon photonic modulators for metro and long-haul coherent systems with performance comparable to lithium niobate.

Tunable laser technologies on silicon photonics were also presented by companies such as Skorpios and Aurrion during the post-deadline sessions.

"Cisco and Alcatel-Lucent presented results on silicon photonic modulators for metro and long-haul coherent systems with performance comparable to lithium niobate."

All those new silicon photonics technologies could eventually become key building blocks of future highly-integrated transceivers.

Pluggable coherent modules will be a big market opportunity and it is all about density and low power consumption.

At the show, Oclaro demonstrated key milestones to bring to market a CFP2 coherent module by mid-2014 while this product is on the roadmap of all other major transceiver vendors.

From Teraxion’s perspective, our recent acquisition of Cogo Optronics Canada for high-speed modulators is directly in line with this market trend at the modules level where performance, size and low power consumption are key requirements.

Paul Brooks, product line manager for high-speed test solutions, JDSU

The growing confidence in second-generation 100 Gig CFP2s was evident at the show. Many companies, including JDSU, demonstrated robust second-generation 100 Gig modules which will drive confidence across the whole 100 Gig ecosystem to allow cost efficient 100 Gig clients. Our ONT CFP2 test solution was well received and we spent a lot of time demonstrating the features that will enable successful CFP2 deployment.

"Many companies are openly discussing 400 Gig and beyond, the bandwidth demand is there but considerable technology challenges need to be address"

One thing enforced at the show is the continued importance of innovation in test and measurement solutions required by our customers as we move to 100 Gig+ systems.

Many companies are openly discussing 400 Gig and beyond, the bandwidth demand is there but considerable technology challenges need to be address. The intellectual horsepower present at the show allows fruitful and engaging discussions on key topics.

See also:

Part 1: Software-defined networking: A network game-changer, click here

Part 2: OFC/NFOEC industry reflections, click here

Part 4: OFC/NFOEC industry reflections, click here

Part 5: OFC/NFEC 2013 industry reflections, click here