Comms chips

Big Bear Promises Picnic

Next Monday (May 20), startup Big Bear Networks is planning to tell the world how its signal processing technology can save carriers millions of dollars.

The timing of the announcement is a little arbitrary. Big Bear doesn't have any products ready yet, and it won't be announcing those until the fourth quarter. Nevertheless, the details of the technology and its applications make interesting reading.

Previously, Big Bear said it was making 10-Gbit/s and 40-Gbit/s optical modules that use signal processing to clean up optical signals (see Big Bear Comes Out of the Woods and Big Bear Hugs $40M). The basic idea is to use electronics -- rather than optics -- to correct for signal impairments that would otherwise cause signals to smudge as they travel down an optical fiber. The electronics can easily be integrated into optical modules.

"Chromatic Dispersion and Polarization Mode Dispersion (PMD) are known issues that, quite frankly have not been solved to date in a way that addresses carriers' capex and opex requirements," says Big Bear's VP of product management Laura Adams. In other words, existing solutions cost too much to buy, install, and maintain, she contends.

Take PMD, for example, she says. Right now, carriers use electrical regenerators to clean up the signal. Not only are regenerators expensive, they are also costly to maintain, requiring an upgrade whenever the speed of the data on the link changes. What's more, when systems (eventually) step up to 40 Gbit/s, many links will not be able to support the higher speed without additional regenerators.

Big Bear's solution to this problem is entirely electronic. The startup claims that it can do PMD compensation using an electronic chip that's placed in the terminal equipment. The chip implements a patent-pending algorithm that corrects both first- and second-order PMD.

The resulting improvement in signal quality makes it possible to do away with some of the regenerators on the link, Adams claims, perhaps even doubling the distance between them. On a 1500km link, that would allow carriers to eliminate two regenerators -- a cost savings in the region of $2 million (based on a 20-channel system running at 10 Gbit/s).

The cost of adding "photronic signal processors," as Big Bear calls its technology, is small potatoes by comparison. It estimates $500 per wavelength, leading to a total cost of $40,000 to compensate the entire 1500km link.

"Cost is a moving target, but we think this is a reasonable estimate," Adams says.

It's worth pointing out that other startups, such as Yafo Networks are using optics to tackle PMD (see Can Yafo Lift Speed Limits?). According to Adams, however, optical compensators will always be more expensive than electronics. She estimates a cost of $12,000 per wavelength -- about 60 times more. In addition, optical compensators take up a lot more space, requiring a complete blade of components per wavelength in Yafo's case.

Yafo was not available to comment.

PMD is just for starters, says Big Bear. Its technology can also correct for other impairments, such as bandwidth limitations at the receiver and self-phase modulation -- an effect that occurs at high optical powers and turns the rounded pulse into a spike (see Nonlinear Effects). And it can help with chromatic dispersion, although it can't do the compensation entirely in electronics, as it can with the other impairments.

For chromatic dispersion, Big Bear uses electronics to control the optics, making the system adaptive to changes in the condition of the optical link. Unlike PMD, which is a constantly changing effect, chromatic dispersion is fixed for a particular link. In the future, however, link lengths could change in systems that use all-optical switching. Even with a fixed link system, adaptive control eliminates the need for an engineer to go out and set the value on the dispersion-compensating module by hand.

It will be interesting to see if Big Bear decides to sell its technology in chip form, or whether it is still intending to integrate chips into optical modules. The company won't say. If the latter, then it looks like it is going to compete head-on with startups like Kodeos Communications Inc., and Optium Corp., which are also developing so-called "intelligent transponders" (see Kodeos Gets Started With $12M and Kalkhoven Opts for Optium).

Kodeos and Optium appear to be further ahead than Big Bear, as both have products out. "We are starting to hear other companies talk about electronic equalizers," Adams acknowledges. "But they only fix a narrow range of impairments. We address the full range of fiber impairments."

— Pauline Rigby, Senior Editor, Light Reading
Be the first to post a comment regarding this story.
Sign In