Customers and Suppliers of Tunable Lasers Report Imminent Deployment Plans
By: Erik Kreifeldt
Near the top of the wish list of carriers that use dense wavelength division multiplexing (DWDM) technology is "tunable lasers." DWDM systems use dozens of different transmitters and receivers that operate at different wavelengths. Keeping spares of the expensive devices in inventory is an unattractive, but critical, proposition, lest one of the devices fail. By tuning to match a DWDM wavelength, carriers can use a spare inventor of a few tunable lasers that could replace any number of devices in a DWDM system, instead of keeping dozens of extra fixed-wavelength devices in stock.
Commercial viability
The challenge for manufacturers is to produce tunable lasers that are as robust as fixed-wavelength devices at a reasonable cost premium. After a few years' worth of work, that goal is coming to fruition. "I think you'll see the first one show up this year," says Jack Wimmer, vice president of network technology and planning at <%=company%>, in response to a query about tunable lasers at the All-Optical Networks conference (Advanced Communication Technologies, Feb. 16-18 2000, Dallas).
Suppliers echo Wimmer's sentiment. Tunable lasers from Kista, Sweden-based Altitun Inc. will carry real traffic by the end of the year, says Basil Garabet, senior vice president for marketing and sales. To bring the devices to commercial fruition, "pricing has been the prohibitive threshold to get through," he notes. To make tunable lasers a commercially viable proposition, Altitun tries to strike a balance between manufacturability and reliability and price. The company expects to offer tunable lasers for a mere 15% premium to comparable fixed wavelength distributed feedback (DFB) devices in about two years. "That opens the door to get used in the fixed part of WDM networks," he says.
Customers
While Altitun has developed its tunable products and geared up manufacturing capacity, the supplier has cultivated a customer base of systems integrators. About two thirds of Altitun's customer base is made up of traditional telecom equipment manufacturers with large research labs in the US, Europe, and Japan, Garabet reports. The other third is comprised of start-up companies. Among the major telco vendors, about half the work is hush-hush, but rapid switching speed is a common theme in device requirements. The switching time parameter implies tunable laser duties beyond sparing, such as optical add/drop multiplexing (OADM).
Citing a trade-off between output power and tuning range, Garabet reports that Altitun is approaching the ability to produce lasers that tunes over the entire optical amplifier C-band and provide 10 dBm output power. Some system integrators seek 5 dBm lasers, while others demand 14 dBm, he notes. For sparing applications, engineers seek out one or two tunable lasers that cover the entire C-band of erbium doped fiber amplifiers.
Sparing applications
Most of the near-term market potential for tunable lasers comes from sparing in long-haul DWDM systems. "For sparing, you need to cover as many channels as possible, but the power has to be close to or at that of a DFB," Garabet explains. Because the laser module is a replacement card, in can't be swapped out for a fixed DFB with a different output power, because the installed system's power budget can't be changed. Designing a tunable laser transmitter module for sparing is not much different than designing it for initial deployment, he notes, because those seeking spare laser modules seek permanent replacements.
Behind long-haul DWDM sparing applications lurks the nascent metro market. "We think power budgets are less of an issue [here]," Garabet says. Wide tunability and cost-effectiveness will take precedence in the metro tunable laser applications, he says. Initial estimates for metro power requirements range from nil to 5 dBm, he says.
OADM applications
Looking beyond sparing applications, Garabet says, "the next wave we see our lasers being used for is for add/drop multiplexers, in conjunction with tunable filters." Tunable lasers that perform nanosecond switching from wavelength to wavelength will enable all-optical network switches and wavelength routing, he predicts.
Moving from sparing to OADM applications requires tunable filters in addition to tunable lasers. Garabet reasons that if you can tune a laser, then you can tune a filter. Altitun, while holding a vested interest in the technology's development, has not entered the tunable filter space. "We have not pursued it with a vengeance," Garabet says. Altitun is banking on the robust DWDM market to push tunable filter technology to mature quickly.
OADM trial
Altitun has participated in a trial of such technology with Norway's Norwegian carrier Telenor AS as part of the European ACTS program. Altitun has also worked with Stanford University at Sprint's network technology lab in California.
Culminating a two-year project, Telenor demonstrates a reconfigurable "switchless" mesh DWDM network based on laser technology from Altitun, Marconi, and NTT. The testbed consists of a five-node, 100-channel network with 50 km and 60 km fiber spans between nodes and 50 GHz channel spacing. The carrier switches wavelengths by varying the frequencies of the transmitter lasers. The network routes optical channels passively rather than relying on conventional electrical or electro-optical switches.
"To our knowledge this is the first time an architecture has been developed that leverages the unique capabilities of tunable lasers," says Björn Broberg, vice president of engineering at Altitun. "We proved that you can build a system with off-the-shelf components from three manufacturers in which the resulting programmable transmitters tune precisely and reliably."
"Clearly tunable lasers are on a fast track," says Evi Zouganeli, project manager at Telenor R&D. "The project confirmed that they lead to elegant, cost-efficient network solutions. We are convinced that the technology will be key to enabling data driven, intelligent optical networks."
Production and design
With feasibility demonstrated, volume production is the next critical component establishing a tunable laser industry. The package, pigtail, and manufacturing processes for Altitun's laser products are the same as for DFB devices, which gives the company the advantage of using established techniques. "We're really going for large-scale production," Garabet says, starting with a facility that has 600 sq. ft. of MOCVD cleanroom space.
Part of the advantage of Altitun's tunable laser approach is that it uses semiconductor lasers with no moving parts, external mirrors, or external cavities, Garabet says. "We believe lasers used in telco applications will not have moving parts and not be temperature tunable," Garabet declares, citing difficulty with complying with a 20-year lifetime, otherwise. He says that mechanical, MEMS-type approaches to tunability are less mature, and may eventually be more applicable to filters than lasers.
Lucent has met with some success with a temperature-controlled tuning approach, but with limited bandwidth, Garabet says. Nortel has worked on a layering approach to the temperature tuning to improve performance, he adds.
Altitun is pursuing two types of tunable laser approaches, distributed Bragg reflector (DBR) and grating-assisted codirectional coupler with rear-sampled reflector (GCSR). Other DBR approaches are under development at Alcatel and JDS Uniphase, Garabet notes. Marconi uses a similar approach to Altitun for a GCSR-type laser, but has so far kept the use of the device internal, he says. "DBRs cannot be tuned widely, but you get high powers," he notes. GCSRs, meanwhile, lag 2 dBm to 3 dBm behind comparable DBRs in output power, but tune over a wider range.
Altitun conducted a joint demonstration with JDS Uniphase at OFC '99, using a JDS Uniphase chip and Altitun control electronics. Altitun differentiates its DBR products by selling them with integrated channel characterization, software, and drive electronics. "It's not a trivial undertaking," Garabet says. "Systems designers may be more interested in transmission than they are characterizing wavelengths of each channel," he surmises.
"We have a high number of device-hours in testing," Garabet notes, citing 120,000 hours. He also notes that tunable lasers are new territory for telecom. "There's no Bellcore standard. We're almost creating a standard for ourselves to start measuring on."