LIGHTCONNECT decided early on not to move into the crowded switch business, preferring instead to focus on components that would support dynamic reconfiguration; a market that the company believes will be significant in future networks.

Organisation
LIGHTCONNECT employs 45 staff of which 25 are engineers. Initial funding, in June 2000, amounted to $8.4 million, and on 9th July 2001 LIGHTCONNECT announced further funding of $15.8 million. Based on its present strategy, LIGHTCONNECT believes current funding will see it "well into 2003".

LIGHTCONNECT opened approximately 5000 square feet of manufacturing space in Newark, California, in May 2001, with about 25 percent currently utilised. Taking into account production shifts, the facility offers the company a factor of eight multiplier in terms of ramping up production.

After launching LIGHTCONNECT, Bloom and co-founder Asif Godil found that a device based on the original Stanford invention lacked some of the properties needed for optical communication applications. Bloom and Godil, now CTO, put their heads together and invented unique new solutions to address issues like wavelength independence and polarisation independence. The company has thus far filed seven patents relating to this "elegantly simple technology" and the design solutions around it. Three patents having been granted to date.

The basic technology of the grating modulator is licensed from Stanford University. Peter Clark noted that the Silicon Light Machines subsidiary of Cypress Semiconductor also has a license for the technology; although he has yet to see Silicon Light Machines in the optical space.

Manufacturing and outsourcing
Due to the nature of LIGHTCONNECT's MEMS devices, manufacturing can be outsourced to CMOS foundries rather than specialised MEMS-type facilities. This offers more choice in terms of support for the product. A further advantage of this approach is that LIGHTCONNECT avoids the costs associated with maintaining a MEMS fabrication facility. The inherent simplicity of the device, requiring lithography of 1.0 micron, means that no new semiconductor processes are needed.

Telcordia testing of the VOAs will have been completed by September 2001 (1209 is complete) and of the DGE by December. Preliminary Telcordia testing has proved to the company's satisfaction that design parameters have been established in the manufactured products, according to Dr. Clark.

VOAs
LIGHTCONNECT continually benchmarks against competitors' data sheets and claims, in the VOA market, to have the smallest and fastest device. Therefore, LIGHTCONNECT argues that in applications where size and speed are critical, it is the industry leader.

In addition, across the full spectrum of optical parameters, from insertion loss to PMD, Dr. Clark believes LIGHTCONNECT either meets the industry standard or betters it, and is on a par with the industry leaders. In terms of reliability, LIGHTCONNECT's VOAs have been tested over several billions of cycles without any measurable change in the optical parameters.

DGE
Dr. Clark noted that there are "very few" competitive DGEs currently advertised. Those that are available tend to have 5 or 8 segments, and are very limited in their ability to shape the gain curve. One of the principal applications of a DGE is to work in conjunction with an EDFA, and in LIGHTCONNECT's case, there are 24 channels, which, Dr Clark argues, offers a far greater ability to customise the gain profile to various EDFAs.

As in the VOA market, Dr. Clark claims the DGE can optically match any other product, and has size, speed, reliability, and power requirements that are not matched by the competition. The enthusiasm and response that LIGHTCONNECT is receiving from customers supports these comments, according to Dr. Clark.

Future products
Dr. Clark also noted that the market for its products has barely taken off, and that system integrators will find applications not currently thought of. According to its product roadmap, LIGHTCONNECT is evaluating the introduction of a VOA with integrated tap detectors.

The next obvious step in the DGE market is a Dynamic Channel Equalizer (DCE) that slices the wavelength region up into 50 GHz channels. LIGHTCONNECT notes that the DCE is a clear extension of its present technology.

With its established engineering and manufacturing capability, Dr. Clark feels that LIGHTCONNECT can bring a modification of an existing product to market in about 3 months. Prototyping a new product normally takes about six months.

Primary applications
LIGHTCONNECT targets long-haul and metro markets with both of its products. The VOA is a component that allows continuous optimisation of the transmitted wavelength spectrum. "As you add or drop wavelengths or channels, there is a need to to continually correct, and any top-level system diagram will show many attenuators that are used for optimisation."

The DGE sits in an amplifier chain as a separate component. However, Dr. Clark expressed the opinion that "Over the next few years, you will see DGE's incorporated into the amplifier module."

Feeling within the company is that LIGHTCONNECT has the ability to be a several hundred million dollar company within a few years, based not just on the first two products, but on a strategy and product roadmap which calls for the frequent introduction of new products. "There is no future in being a two-product company, you have to have a roadmap and a technology that supports the evolution of your roadmap", according to Dr. Clark.

He noted that, in common with the general industry, he was concerned about current market conditions, but said "During this time we will work closely with our customers, who are listening to their customers, to design next generation products that will be well positioned when the upturn begins." In Dr. Clark's view, based on reports from his own customers, the market for systems will not turn upwards until the second half of 2002, though, as a component supplier, "LIGHTCONNECT has an advantage on that timeframe".

Competition
In the VOA space, LIGHTCONNECT views JDS Uniphase and Corning as prime competitors, along with smaller companies like MOEC and Digilens, but notes that they all have different technology.

In the DGE market, which LIGHTCONNECT considers "a different ball game", Novera and Chorum are cited as key competitors. "But the only thing that these companies have that we don't is a longer time in the market."

Competitive edge
Dr. Clark summed up LIGHTCONNECT's edge as "better performance and better value." LIGHTCONNECT's products are provided in smaller packages, a huge advantage where optical component footprints are important, and Dr. Clark believes he can offer customers a very competitive price. LIGHTCONNECT concedes however that the VOA market is much more crowded.

Dr Clark recognises that his biggest problem is getting "face time" and breaking through the barrier where potential customers see six or seven suppliers with VOAs. As this type of product becomes more mature, and the market begins to set more aggressive prices, he foresees LIGHTCONNECT's strong participation.

Dr. Clark made the point that despite his company being in stealth mode for so long, products are being shipped to publicly traded companies, and in every case, despite the fact that they started looking at just one of LIGHTCONNECT's products, all are now looking at both.

Route to market
LIGHTCONNECT's present strategy is to go to market direct, though coming from Hitachi Semiconductor where he was President and CEO, Dr. Clark is very aware of other channels and plans to evaluate alternatives "according to product and time". LIGHTCONNECT is already talking to customers outside of the U.S.