Free-space optics 'available' for the last mile?

By Meghan Fuller

Recent announcements from Lucent Technologies, TeraBeam Networks, Air-Fiber Inc., and Nortel Networks have generated renewed interest in free-space optics, particularly as a technique for last-mile telecommunications. However, the technology may have to prove its "availability" before carriers fully embrace it.

Much as microwave radios send RF signals through the air from transmitter to receiver, free-space optical equipment sends light from origination to destination without the use of fiber. In theory, free-space optical networking could be the Superman of data-transmission technologies. Such systems can go where economics or terrain may prevent the use of traditional fiber in populated urban areas and campus environments and across highways and rivers. They promise the high-bandwidth-carrying capacity and cost-curves of optical fiber and are more secure than point-to-point microwave and Local Multipoint Distribution Service. And they don't require spectrum regulated by the Federal Communications Commission.

Such promise has attracted the attention of some of the biggest companies in the optical-networking field. For example, AirFiber (San Diego), manufacturer of the OptiMesh system introduced at SUPERCOMM last month, has announced an OEM agreement with Nortel Networks (Ontario, Canada), granting Nortel the right to market the OptiMesh equipment to its customers worldwide.

Unlike traditional point-to-point free-space optics, the OptiMesh features a mesh architecture that extends like an umbrella over an urban or campus environment, providing bandwidth traffic to each of the buildings touched by the mesh. A node (which features four optical links and an ATM switch) is placed on the rooftops of buildings with existing fiber and on adjacent buildings without fiber. The nodes are then connected in a mesh configuration via line-of-sight optical links. Every building in the network has multiple links going into it, which gives AirFiber the ability to reroute traffic for temporary blockages or equipment failures.

Shortly before SUPERCOMM, Tera Beam (Seattle) and Lucent (Murray Hill, NJ) announced the joint establishment of a new company, to be known as TeraBeam Internet Systems. The company's IP-based Fiberless Optics point-to-multipoint approach includes a transmitter and receiver the size of a small satellite dish, which is mounted in an office window. These dishes beam traffic to and from a base station mounted on a service providers' building.

In July 1999, Lucent introduced the WaveStar OpticAir product line, which promised to deliver information through the air at a rate of 2.5 Gbits/sec over distances ranging from 4.4 to 5 km. Why then would Lucent partner with startup TeraBeam if it already has the technology in-house? According to Frank Briamonte of Lucent's optical media-relations department, the OpticAir line, once heralded for its innovations in WDM and DWDM, is no longer coming to market. Briamonte was quick to note that the decision to partner with TeraBeam was not the result of a failure on the part of Lucent. Rather, the company has decided that such a partnership will be mutually advantageous; Lucent will reap the benefits of a point-to-multipoint system (the OpticAir was point-to-point), and TeraBeam will benefit from Lucent's WDM and DWDM technology.

Eric Korevaar, CEO of AstroTerra Corp. (San Diego), whose company provided telescope transceivers for Lucent's successful deployment of a 10-Gbit/sec laser-communication system in February 1999, asserts, "The joint venture that Lucent formed with TeraBeam Networks came as a surprise to us as well as, I think, to the people we were dealing with at Lucent. What it looks like is that networks such as TeraBeam is proposing would also have good use for equipment such as the OpticAir equipment to connect the hubs together."

Just how much of Lucent's OpticAir technology will appear in TeraBeam's Fiberless Optics is unknown; at press time, company officials were unavailable for comment.

The secrecy of companies like Lucent and TeraBeam may only feed the doubts of carriers that have been reluctant to field free-space technology in the past. While the reliability of the systems has not been much of a hurdle to deployment, availability-the percentage of time the equipment will operate due to weather/atmospheric conditions-has proved a difficult issue to overcome.

Admits Korevaar, "If the costs were equal, fiber-optic cable connection is probably a more favorable solution than installing a laser-communication connection, because there are some weather conditions-once you get out past a couple hundred meters-where you are not going to have 100% availability from the laser links."

Costs may not be equal, however, especially considering the price-in terms of money and time-of trenching city roads to deploy fiber. The demand for a feasible alternative such as free-space optics is therefore growing. The problem now is how to overcome these availability obstacles.

Critical to the success of a free-space optical network is a direct, uninterrupted line of sight between transmitter and receiver. Rain, snow, smoke, atmospheric scintillation, and building movement resulting from solar and wind loading can block transmission between laser links. The big obstacle, however, is fog, which is a serious attenuator of light at longer transmission distances.

For last-mile telecommunications offerings such as those proposed by AirFiber and TeraBeam, the catch phrase is "four nines" or even "five nines," meaning that equipment must function 99.99 or 99.999% of the time. According to Korevaar, there are three ways vendors and users can achieve this goal: (1) Deploy networks in a place where there is never any bad weather (obviously, not a feasible option); (2) keep laser links very short, which is the route AirFiber has chosen; and (3) back up a solution with some other kind of solution.

Each link in AirFiber's networks has five-nines availability, claims the company, due in part to shortened link lengths and the company's patented system, Automatic Tracking. As part of this effort, AirFiber creates a fog profile for each city in which its equipment will be deployed and, based on frequency and severity of fog, calculates the "five-nines" link range specific to that area.

According to Janet McVeigh, AirFiber's vice president of marketing, most cities have a five-nines range between 180-200 m and 500 m.

TeraBeam also claims to have overcome the availability issue-though how it does so remains unclear. The company asserts that it creates a footprint over dense metropolitan areas where its systems are deployed, and this footprint accounts for the weather conditions unique to that location.

TeraBeam is not alone in its ambiguity on this subject. Lisa Pierce, vice president of the telecommunications research practice at Giga Information Group (Cambridge, MA), says "When you ask questions about exactly how much environmental stuff [the equipment] can endure in terms of rain and fog, [manufacturers and service providers of such solutions] don't get very exact. Obviously, that's a concern because companies cannot control the weather."

So then, are Lucent and Nortel getting in on the ground floor of something with real staying power, or will the availability obstacle prove too difficult to overcome? McVeigh is confident about her company's product: "We're looking at a technology here that isn't just going to serve the odd building; it's not a niche technology. It's the kind of technology that can take high-bandwidth services from 3-5% of your buildings up to 50-60% of your buildings."

Pierce, however, is reluctant to accept the technology as it now exists. "I understand using [the technology] for an event-a football game and that kind of thing-but for regular day in, day out business-type stuff, I think we need to know a lot more than we do right now," she says.

While TeraBeam and AirFiber, with their point-to-multipoint and mesh solutions, are thought by some to be well ahead of others deploying traditional point-to-point systems, the technology is some time from widespread commercial deployment. TeraBeam expects to roll out service to 100 major business centers over the next four years, a fact that may speak volumes about where the market is right now.

Says Pierce, "Why is it going to take [TeraBeam and Lucent] four years, or three and a half years to get to a hundred cities? If you're going to do it for day in, day out business applications, you have to roll it out fast." Thus, while free-space optical networking has potential, just how much potential remains to be seen.

Meghan Fuller is assistant editor of Integrated Communications Design magazine, Lightwave's sister publication.