Broadband networks, once the province of behemoth institutions, are looking to migrate toward residential areas in the hopes of carrying a boatload of entertainment and communications services to the doorsteps of every home. Trouble is, wangling that coveted invitation hinges upon the ability of cable and telecommunications companies to go easy on consumer pocketbooks.
Solving this low-cost, high-bandwidth puzzle is just what companies like LightPath Technologies and Invention Machine Corp. propose to do. Through their joint venture, LightChip, the two companies are developing a wave-division multiplexing (WDM) scheme that will use multiple optical lenses fused together into the shape of a microscope slide that is curved to control a light beam's path.
"We need a common technology that will be cheap enough so 100 million homes can get these services in the manner in which they're already accustomed - low monthly fees," said Paul Dempewolf, director of optical electronics product development for the Albuquerque, New Mexico-based LightPath.
The refinement of WDM devices is the focus of several industry efforts to help a variety of data travel simultaneously along fiber-optic cable at varying speeds. Already, WDM is used for larger networks to carry data long distances. But the devices and supporting infrastructure for these "long-haul" systems are expensive - roughly US$250,000, Dempewolf estimates. For data to travel along these distances without loss, companies have to use stronger, more expensive lasers - along with cooling systems, shielding to "ruggedize" the fiber, and amplifiers to boost the signal at the end of its journey. Add all of these expenses together, and you have a formula that most telecommunications and cable companies won't touch, Dempewolf said.
"At $8,000 to $9,000 for a splitter, most telecommunications companies won't bother with this technology," said Wayne Knox, distinguished member of the technology team at Lucent Technology's Bell Labs, which earlier this year introduced a chirped-pulse WDM technology that is also intended to bring high bandwidth to homes and businesses - eventually.
And WDM is going to trickle down to smaller and smaller networks as bottlenecks get pushed from the major backbones to the central office networks of the telephone companies and onto metro networks, says Bill Magill, vice president of the research group at Montgomery Securities. "There's a huge drive to make low-cost WDM, but no one quite knows how to do this yet," he said.
The focus of LightChip's effort is to perfect its patented Gradium lenses for use in telecommunications applications. The curved lenses are bent in varying angles, calculated to diffraction indexes, to divide a laser beam that is directed along the length of the slide.
At close range - at home or two blocks away - Dempewolf says the LightChip WDM transmitter would allow telecommunications companies to relax the specifications for other parts of the data transmission equation. So companies could use cheaper lasers and skip having to use cooling devices and shielding because data won't have to travel as far and tax the system as much as a long-haul setup would.
The target pricetag Dempewolf hopes to hit is around $200 per node - or less.
Price is the most intriguing part of the proposal to Magill. "They're taking an older technology that's well understood [gradient glass] and putting it to this new use - focusing on ways to make it cheaper," he said. "But then, who isn't trying to [make low-cost WDM]?"
Who isn't indeed. In his travels as a partner for the Massachusetts-based venture capital firm Battery Ventures, Todd Dagres is aware of at least three companies around the world that are working to bring WDM to local loop communications networks such as those currently maintained by the Baby Bells.
But this development is something that is a ways off from being reality. While Dempewolf waxed confident about reaching the price point, he hesitated when asked about the timeline for testing and installation of the technology.
Dagres, who is watching this local network WDM market evolve, could scratch out a timeline of sorts. "The [Baby Bells] will take two years to evaluate and test the technology once it's been developed. They won't deploy it until the third year after the technology is out. So this is at least five years away," he said.
