Metro DWDM: Not the Only Game in Town
The startup is finalizing a $50 million funding round that will enable it to roll out an alternative technology: optical code-division multiple access (OCDMA). CodeStream is counting on it to give metro DWDM (dense wavelength-division multiplexing) vendors quite a run for their money, particularly as the economics of deploying their gear are still in a shambles (see Report Forecasts Metro Winners, Losers ).
On a box-by-box basis, CodeStream's costs will be comparable to those of metro DWDM, according to Gordon Werner, the startup's president and CEO. But installation, maintenance, and provisioning costs will be substantially lower, he claims.
OCDMA is compatible with DWDM, in the sense that one system can be interfaced to the other, for example, where the metro network links to the long-distance systems. But technologically it's so radically different from anything else out there, that gaining acceptance from metro carriers could be a significant hurdle.
CodeStream, of course, reckons it will win carriers over. To see why, let's take a closer look at the technology involved.
OCDMA is an optical adaptation of a technique that's already used to increase the capacity of mobile phone systems. Each channel is assigned a unique code -- basically a combination of frequency components -- which makes it possible to overlay multiple channels in the same bandwidth space. CodeStream likes to call it "bar-coding the light".
To do this, an OCDMA system requires a broadband light source, spectral filters to apply the code, and a modulator to add the signal to the coded channel. Since all the channels occupy the same bandwidth, one light source is sufficient for all the channels (in contrast to DWDM, which requires a high-precision laser per channel). At the detection end, filters are required to decode the channel and extract its data. Intermediate points on the network can be added at any time simply by installing an optical tap to draw a little power from the fiber, which contains information from all the channels.
Right now, CodeStream claims to have a working system with 15 channels in its laboratory. The broadband light source is an erbium doped fiber amplifier (EDFA) pumped by a light-emitting diode. This provides light at around 1550 nanometers with a bandwidth of 28nm (limited by the gain spectrum of the EDFA). Fiber Bragg gratings are used to impress the codes onto the various channels.
However, this is not likely to represent the final configuration of CodeStream's product. "Typically, in long-haul systems circuits stay static, but in the metro space they need to be reprovisioned on a routine basis," explains Werner. Fixed filters like Bragg gratings can't do dynamic reprovisioning, he says, so CodeStream is working hard on an alternative, based on micro-electro-mechanical systems (MEMS).
Exactly how MEMS can do complicated filtering is not clear. All Werner will say is that the tilt of the mirrors can be altered to select different frequencies of light. The fact that CodeStream's director of advanced technology, Dr. Gregory Magel, is a world-renowned authority on MEMS implies that the success of this idea is pretty central to the company's strategy. There's one other factor that could have a strong influence on the startup's future -- the price of EDFAs and other optical amplifiers. Right now, EDFAs are considered to be too expensive for the metro network. Strangely, because of its architecture, an OCDMA net could require dozens of these devices. For a start, each source requires an EDFA, which has to supply power to all of the channels. In addition, amplifiers will be required to boost optical power every 20 kilometers or so, to make up for the losses incurred when the signal is tapped. CodeStream is banking on the cost of EDFAs coming down significantly in the next 12 to 18 months, by which time it hopes to be in volume manufacturing.
CodeStream says it made some technology announcements a couple of years ago, but these proved to be hopelessly premature. Since then the company has kept its nose to the grindstone. "We made more technical progress in the last six to nine months than in the previous three years," claims Werner.
Founded in 1996, CodeStream previously received funding of $20 million, mainly from SpaceVest, Kline Hawkes, and Capital Communication. In June 2000, the startup was acquired by CodeStream Holdings Inc. (Nasdaq: COHO). Since then, it has raised money through private placements of common stock. As a result, 98% of CodeStream's stock is restricted; the remainder is traded over the counter.
-- Pauline Rigby, senior editor, Light Reading http://www.lightreading.com