by Stephen Hardy
While the twin pillars of network flexibility-tunable lasers and reconfigurable optical/add drop multiplexers-as well as FTTH received their expected due, new technology announcements and demonstrations in other market niches at this year’s OFC/NFOEC in Anaheim, CA, underscored the fact that network planners have regained interest in product innovation rather than focusing strictly on lower costs. The inter-related goals of higher transmission speeds and extended reach, as well as the quest for smaller transceivers, provided plenty of conversation within the Anaheim Convention Center March 26-29.
Carriers continue to seek a reduction in equipment costs-but that’s no longer the only thing that interests them, say sources. “The recent emphasis is more on dollars spent per bit transmitted-or maybe another way to say it is the cost of ownership of the network is actually where the focus is. So that doesn’t necessarily mean lower-cost components,” asserted Giovanni Barbarossa, chief technology officer at Avanex (www.avanex.com), in an interview before the show. “From my perspective, it’s more about new components that provide the network operators lower capex and lower opex opportunities to manage their networks.”
For example, carriers face the challenge of how to deal with rising bandwidth demands. This problem is particularly acute on major routes where service providers face fibre and 10 Gbit/s wavelength exhaust. Rather than install new racks of 10 Gbit/s equipment or new fibre, carriers have expressed an interest in 40 Gbit/s transmission; Verizon, in fact, announced during the show that it would transmit traffic at 40 Gbit/s between its New York and Washington, DC, hubs. Vik Saxena, senior director, network architecture within the Corporate CTO Office of U.S. cable multiple systems operator (MSO) Comcast (www.comcast.com), suggested at the OSA/Lightwave Executive Forum that his company has plans to move to 40 Gbit/s as well, particularly within the company’s core and regional/metro distribution networks.
To make such a jump both capex and opex friendly, service providers will want to avoid laying new fibre or repositioning (or adding) amplifier and regeneration sites. However, the heightened effects of dispersion-both chromatic and polarisation mode dispersion (PMD)-at 40 Gbit/s present obstacles to the reuse of installed resources. Meanwhile, as a workshop on PMD held March 26 pointed out, dispersion also can challenge 10 Gbit/s transmission. Dispersion management therefore has become a hot topic of research, the results of which were on display in Anaheim.
For example, Avanex, Civcom (www.civcom.com), and OFS (www.ofs.com) displayed adaptable dispersion compensation modules on the OFC/NFOEC show floor. Avanex’s PowerShaper 3200 remotely tunable dispersion compensator is designed to provide 100% slope across all C-band channels simultaneously to aid 10 and 40 Gbit/s transmission. The module can be integrated into the mid-stage of an EDFA or a DWDM demultiplexer. Civcom added a multiport option to its Manageable Dispersion Compensation (M-DCM) line, which provides a variety of in-line and pre- and post-compensation for single- and multiple-channel operation. The company introduced dual-port and “multiple-port” versions of the product, the latter of which can scale up to 16 ports. OFS, meanwhile, placed a new spin on dispersion compensating fibre with its Reconfigurable Dispersion Slope Compensating Module (R-DSCM). The module contains a set of switches that enable the network operator to switch among four different dispersion compensating fibres within the same module.
Researchers also have focused on alternatives to the standard non-return-to-zero (NRZ) modulation format as a means to combat dispersion and enable either longer reach at 10 Gbit/s or maintain reach at higher speeds. OFC/NFOEC attendees heard discussion of a variety of such formats, including optical duobinary, differential phase-shift keying (DPSK), and differential quadrature phase-shift keying (DQPSK); Mintera (www.mintera.com), meanwhile, touted another alternative to DPSK, partial DPSK. Visitors to the show could find examples of all of these technologies, either in the form of modulators, compatible ICs, or transceivers. DQPSK was frequently described as enticing but expensive, which should leave the door open for duobinary; transceivers using this technology have already been announced by such companies as Civcom, Kodeos, Essex (www.essex.com), and Kailight.
A series of acquisitions announced at the start of the show underscored the attention this space has attracted. Finisar (www.finisar.com) announced it had agreed to purchase Kodeos and AZNA LLC; the latter’s chirp-managed laser technology provides benefits similar to alternative modulation schemes for high-speed transmission. (It should be noted that both companies also give Finisar tunable laser capabilities, a factor that undoubtedly held at least as much attraction as the dispersion-resistance expertise.) Optium countered by picking up Kailight; the new acquisition announced capabilities in duobinary, DPSK, and 40 Gbit/s NRZ two weeks before OFC/NFOEC opened.
Companies also discussed advances on the receive end of the equation, with coherent receiver/detector technology suddenly raising its profile. Discovery Semiconductor (www.chipsat.com) showcased its Kitty Hawk rack-mounted coherent receiver system. Coherent receivers are able to extract signal information from more phases of a communication signal than standard detectors, said personnel at the Discovery Semiconductor stand. This capability matches well with the alternative modulation formats that transmit signal information in multiple phases. The company currently positions the Kitty Hawk as a 10 Gbit/s instrumentation platform. However, Abhay Joshi, Discovery’s president, revealed that his company plans to repackage the technology into a format that could be deployed more easily within networks.
A pair of demonstrations reported in post-deadline papers underscored the potential of this technology. In the first paper, researchers at CoreOptics (www.coreoptics.com), Siemens Networks/PSE (www.siemens.com), and the University of Eindhoven reported that they had combined coherent technology with DQPSK to transmit 10 wavelengths spaced at 50 GHz over 2,375 km at 111 Gbit/s. Meanwhile, Alcatel-Lucent (www.alcatel-lucent.com) and the University of Parma reported they had paired coherent detection with QPSK to transmit 40 Gbit/s signals over a 4,080 km link with a high degree of emulated PMD.
Small size, big impact
The drive to develop SFP+ transceivers also attracted a lot of attention at OFC/NFOEC. This transceiver format, which is still undergoing standardisation, aims to provide some of the reach and other attributes that larger form factors can accommodate while maintaining a pluggable package size smaller than an XFP.
Several companies demonstrated SFP+ modules, including Avago Technologies (www.avagotech.com), EMCORE (www.emcore.com), Finisar, Opnext (www.opnext.com), Picolight (which JDSU recently purchased), and Sumitomo Electric Industries/ExceLight (www.sei.co.jp/index.en.html). Although Sumitomo predicted a final multisource agreement for the module would not be completed until this quarter, the company said its SFP+ offerings were already available for shipping in production quantities.
The SFP+ will find a wide variety of applications, including 10GBase-LRM. Because the electronic dispersion compensation (EDC) technology necessary to push 10 Gbit/s over at least 220 m of 62.5-µm multimode fibre will not reside in the SFP+ module, such transceivers compatible with this PMD have followed closely behind those in larger form factors such as X2-most of which also made their debuts at the show. The EDC responsibility for SFP+ sits squarely with the IC vendors, who must come up with different functional combinations for the SFP+ than they did for XENPAK, X2, and XFP. Chip companies displayed their willingness to step up, with AMCC (www.amcc.com), Aeluros (www.aeluros.com), ClariPhy Communications (www.clariphy.com), and Vitesse Semiconductor (the reputed big winner in the first round of X2 designs; www.vitesse.com) all discussing their plans. Scintera Networks (www.scintera.com), one of the early players in the space, was somewhat cagey about its plans for SFP+, but the company highlighted new capabilities in its stand.
Finally, work progressed on bringing tunability to pluggable modules. Several R&D efforts have sought to integrate the tunable laser with a Mach-Zehnder modulator in the same wafer. The resultant device, called an ILMZ (for “integrated laser and Mach-Zehnder”), was a topic of discussion at the stands of Bookham (www.bookham.com) and JDSU (www.jdsu.com).