Tunable lasers, the most highly publicized tunable components, potentially solve a logistics headache. The advent of wave-division multiplexing (WDM) allows multiple wavelengths of light to share the same fiber-optic cable. But it takes one laser to produce each of those wavelengths, and as the power of WDM has grown to 160 wavelengths per fiber, the collection of lasers needed has become unwieldy.

"Each one has its own product code, and each one has to be managed line by line," said Kevin Affolter, director of marketing for tunable-laser manufacturer Agility Communications Inc. (Santa Barbara, Calif.). "If you can go from, say, 192 product codes to just one or two, you don’t have to worry about forecasting the product mix." But companies aren’t willing to pay much of a premium to eliminate the inventory.

"Tunability doesn’t drive any architectures," said Rohit Sharma, chief technical officer of systems vendor ONI Systems Corp. (San Jose, Calif.). "As a systems designer, I may agree to pay a 10 percent to 20 percent premium, but I certainly wouldn’t pay a 100 percent premium," which is what some tunables are demanding.

In fact, price is likely to be the stumbling block as tunable lasers emerge. The problem is that some tunable lasers have deviated too far from the structure of a regular laser. "You would have to achieve tunability with a very simple structure that doesn’t deviate far from DFBs [distributed-feedback lasers] so you can get to high-yield, high-volume manufacturing," Sharma said. The price would be acceptable if the applications were more exotic, but that won’t be the case for tunable lasers any time soon. "The real sweet spot for a lot of the tunability will be in applications that are not on the market yet," said Mike Arden, an analyst with KMI Corp. (Providence, R.I.).

Tunable-laser makers also will need to band together in multiservice agreements (MSAs) that will dictate physical package sizes for their parts. An MSA would ensure that board designers wouldn’t be forced to use a single vendor’s laser.

Without MSAs, "it just makes it that much more of a risk for us, because you’re betting on that company being successful," Sharma said. "A couple of ways of developing tunables will win out. The [manufacturers] may use different technologies, but they have to agree on one package."

Swedish startup Altitun AB, now owned by ADC, was among the first vendors to announce its tunable-laser plans and has an MSA agreement with Agility. Nortel Networks appears to be a front-runner as well after its acquisition of CoreTek; Nortel in March announced it had begun sampling its tunable lasers for 40-Gbit/second transmission.

More than a dozen other startups have been developing tunable lasers, but none has hit on the right formula, according to Sharma. "If you take all of the startups and established players together, collectively they’ve achieved the right things," Sharma said. But no single company has hit them all. "There isn’t an easy, obvious winner out there."

On the receiving end of a tunable laser is a tunable filter to catch particular wavelengths of light. In some cases, companies are developing lasers and filters simultaneously. Many of those companies are building the devices out of microelectromechanical systems (MEMS). Specifically, they’re using rows of pop-up mirrors to deflect the light.

"A number of [MEMS] companies that had indicated they were entering this space announced tunable lasers and tunable filters," said Marlene Bourne, an analyst covering MEMS for Cahners In-Stat Group. Some companies are developing lasers, filters, variable optical attenuators and MEMS cross-connects, hoping to gain some economies of scale by developing all four. "It kind of illustrates how the use of the technology has expanded to a system-level solution. Over a year’s time, three other product areas have been launched, and yet people are still fixated on switching." Bourne predicts a market of $67 million for the four MEMS segments combined this year, rising to $2.3 billion in 2005.