By Wes Simonds

EtherLinx’s wireless LAN technology is in a class by itself… or is it? The company’s offerings are broadband wireless solutions based on modified wireless LAN (WLAN) products. The revised products reportedly extend maximum range by several orders of magnitude — from several hundred feet up to fifty miles.

In this way, the company promises wide area networking (WAN) at far lower costs than previously considered possible, while retaining the raw throughput (roughly two Mbps) commonly associated with today’s 802.11b offerings.

Of course, such a radical reinterpretation of WLAN limits is bound to generate a skeptical response. At a time of corporate scandals from companies as well regarded as Enron and WorldCom, the appeal of EtherLinx’s solution carries a certain stigma — simply put, it sounds too good to be true. That Silicon Valley venture capital firms held a similar opinion was almost certainly the reason EtherLinx was unable originally to raise funds for development.

In addition, EtherLinx has yet to formally release a product; at the time of writing, the company is still in trial phases for product evaluation and development.

Nonetheless, it might just be that EtherLinx’s solution is nothing short of a revolutionary step in the development of last-mile broadband wireless solutions.

Out of the Garage

EtherLinx’s origins have a humble cast. Like Silicon Valley successes Hewlett -Packard and Apple Computer, EtherLinx too appears to have been born in a garage. Likewise, EtherLinx is the technological brainchild of two savvy engineers who decided to take advantage of a largely ignored market opportunity. Layne Holt and John Furrier, the company’s cofounders, appear to have the credentials required to make a go of such a business plan.

Holt’s background is essentially technical. The star bullet on his resume is his stint as president of Data System Services, where he codeveloped, designed and manufactured the proprietary wireless products used by EtherLinx’s solutions.

Furrier has a technical background as well, including a BS in computer science from Northeastern and experience at both IBM and Hewlett-Packard. EtherLinx’s mission plan is simple enough. The company is geared to provide broadband wireless access at lower cost, with faster deployment times and at greater range than has yet been possible.

From LAN to MAN

Impressed by the rapid deployment of Wi-Fi in corporate America, Furrier and Holt realized that what today’s networking arena wanted was a Wi-Fi solution that had a range of many miles (see page 18 for details on 802.11’s range limits). What had been Wi-Fi hotspots, in this scenario, would expand by way of the company’s proprietary solutions to cover whole cities.

Some cities, such as San Francisco, are less than 10 miles from end to end. A
single EtherLinx broadcasting solution, if it yields the range benefits envisioned,
could in theory cover all of that territory.
Thus, EtherLinx was born. And toward that end, the company has made a hefty
media splash. A recent article on the company in the New York Times drew the
attention of many in the broadband wireless industry, as well as a certain
amount of qualified skepticism.

For instance, consider this passage from the company’s promotional literature:

EtherLinx extends usable link distances of off-the-shelf 802.11 radios, using FCC-approved techniques. Traditional 802.11 radios have usable link distances of only a few hundred feet. EtherLinx is able to reprogram the firmware/software to get to usable link distances of 10 to 50 miles, depending on antenna type.

Intuitively, it might seem that the physical range limitations of an 802.11x
broadcast can ’t be multiplied by software alone. EtherLinx Cofounder John
Furrier explained this apparent magic.

“EtherLinx takes standard 802.11b hardware and reflashes the hardware with our proprietary software, ” he said. “In essence, we change the timing of the standard to allow for longer ranges. The tradeoff we made for the longer range was speed. Our solution is 2 Mbps. [We also] have the capability to introduce a 10 Mbps version that only goes 2 miles with our proprietary technology.”

This firmware, known as Smart Spectrum, appears to be the heart of the
EtherLinx solution. Based on some two million lines of code, and generated over
a development period of five years, Smart Spectrum is clearly, according to
Furrier, responsible for much of the optimization involved.

Why WLAN vendors have been unable to develop this type of optimization themselves isn’t wholly apparent, although the complexity implicit in two million lines of code is remarkable. One is left wondering how such an enormous project compiles down to a size suitable for the physical memory limitations involved in traditional Wi-Fi firmware.

Other questions remain unanswered too. It seems that a contracted source of hardware (that is, a WLAN vendor) would need to participate with EtherLinx to develop the solution as described. EtherLinx isn’t a WLAN solution provider in its own right. And, as of this writing, there is no reference to such a source in EtherLinx’s promotional material or Web site.

At this point, there’s little outside opinion available concerning the viability of EtherLinx’s technical solution. Robert Fitzgerald, CEO at Young Design Inc. (YDI), said that YDI has extended the range of its wireless systems in a different way.

“YDI has accomplished real -world links in excess of sixty miles ... but that’s using high-gain amplifiers and high-gain antennas, Fitzgerald said. “We know that much distance is possible, but that’s due to physical hardware upgrades, not software/firmware upgrades.”

Antenna vs. Firmware

Precisely how Smart Spectrum functions is open to debate. Is custom software even necessary? There have been many media reports of basic antenna-based modifications of traditional WLAN products — no custom software involved — that yield vastly improved WLAN range (over 10 miles in some cases). And EtherLinx’s firmware modification does appear to be paired, in its deployed version, with such an antenna, which is certainly not of the type typically associated with typical WLAN products.

But what percentage of the extended range is due to the antenna, and what percentage is due to the firmware, remains unknowable at this time.

“The WLAN products really were not designed for extended ranges and if they are being driven by special modified antennas, then that is one approach,” Furrier said. “We don’t think that scales well, and the simple hack still has the security problem.”

Security

The security problem is certainly worth considering. No aspect of WLAN technology received more press coverage over the course of the last year than the fact that its integrated security, Wired Equivalent Privacy (WEP), has been cracked. WEP was supposed to provide security in wireless LANs comparable with that traditionally found in cabled LANs. WEP is essentially insecure in its base form — that is, when not supplemented by third-party products from companies such as Fortress and NetMotion.

If the range of a WLAN is extended, security concerns escalate as well. For instance, under normal conditions, a WLAN can be tapped by a hacker from a company’s parking lot. In contrast with that, an EtherLinx-boosted WLAN could theoretically be tapped from different zip codes (assuming EtherLinx’s claims on the system ’s range are accurate).

Does EtherLinx offer any special security refinements?

Furrier said that EtherLinx isn’t bothered by the security problems of 802.11b.

The company’s literature, which addresses this matter in more detail, says that dedicated spread-spectrum techniques are used to address the question of security. EtherLinx technology operates in the 2.4 GHz, 5.2 GHz and 5.8 GHz bands, but not in the same way as traditional 802.11b and 802.11a products.

According to EtherLinx, the generated signal appears to be “background noise” to anyone attempting to listen to the transmission. This approach has an additional benefit in that frequency saturation, a concern particularly for 2.4 GHz products, is not an issue. Finally, the company “can layer industry standard encryption techniques on top of its proprietary data transmission technology” for additional security.

Deployment Costs

Of course, deployment costs are as critical as robust security, both to the carriers and to their customers, who ultimately bear the expense of the deployment in the form of fees. EtherLinx’s position on this topic is that because they are reworking low-end, off-the-shelf WLAN products, as opposed to developing and supporting comparatively expensive alternatives, deployment costs are exceptionally low.

How low? The company cites figures of under $100 per customer for the hardware itself. In most cases, truck rolls aren ’t even necessary, according to EtherLinx, meaning that the customer can install the solution directly. What’s more, prices for the basic hardware are expected to fall as the number of vendors increases and competition for market share begins. These savings can be passed on to the user. Thus, it’s possible that even $100 could become an inflated estimate in a relatively short time.

Such a radical shift in the use of Wi-Fi technology calls forth a few questions on how it might be adapted. For instance, could a home user modify an existing WLAN system to work with EtherLinx’s networks? Could the user purchase the firmware and upgrade WLAN products without having to buy new hardware with EtherLinx code preinstalled?

The prospect of updating a WLAN with EtherLinx technology appears to be something for the future. Right now, the company has its hands full.

“We can update a standard card. We haven’t seen that demand yet because we have been flooded with last-mile business opportunities,” Furrier said.

Then there ’s the question of the early reaction of trial customers, currently limited to the Bay Area in the Campbell (South Bay) and Oakland (North Bay) regions. Are they satisfied with EtherLinx’s performance and reliability?

It’s a point worth pursuing since weather (specifically rain fade, which occurs when floating water droplets act to refract and thus slow down network throughput) is a serious problem with other broadband wireless technologies.

One might reasonably expect San Francisco’s legendary fog to be an issue at the nearby Oakland deployment, for instance, more than in the Campbell area, which generally has clearer weather.

But Furrier says it’s not the case. “[We’ve gotten] overwhelming response both in just fan mail and actual requests for service. We haven’t seen any problems with weather or fog in particular.”

Appeal to the Grassroots

Grassroots WLANs could play an interesting role in EtherLinx’s future.

These ad hoc networks are becoming popular in areas such as Seattle, San Francisco, Austin, Texas and New York City. They generally involve users who contract for individual broadband access over cable/DSL and subsequently offer it up to an entire community for free over a WLAN bridge. Thus, a large percentage of people within a given radius would have a full range of Internet services for essentially no charge, given ownership of standard 802.11x hardware. Individually, these “bandwidth Samaritans” would have little impact, but they are organizing themselves so as to make Internet access a commodity.

Naturally, the carriers are unhappy about this developing trend, because it represents a significant loss of potential revenue. Every one of the free, grassroots WLAN users might otherwise have been a carrier data customer. For this reason, some providers such as AT&T and Time/Warner are reported to be cracking down on grassroots WLANs and trying to stall their growth by prosecuting offenders.

On the side of the carriers is the fact that a grassroots WLAN is only as effective as its coverage, and its coverage is generally rather sparse due to the range limits of Wi-Fi.

With EtherLinx technologies, however, those limits would in theory vanish. One could therefore make the argument that EtherLinx is a possible threat to carriers as well as a possible solution.

Furrier has considered these issues as well and seems to fall on the side of the grassroots WLANs.

“Any technology that spreads more broadband capability in a mass way will help the grassroots efforts,” he said. “We are the biggest fans of Wi-Fi. In terms of free vs. paid, we see a healthy balance between both, with an emphasis on more free bandwidth. In terms of free, we envision a day that our technology will create the equivalent of a backbone ‘interstate highway system’ [of] free travel and backhaul. That is the future model, and we are ready to embrace it and drive toward that. As an example, where I used to live in Massachusetts, they had Interstate 95 (free) and Mass Pike (paid) — we see the same for the bandwidth business.”

The word “free” in this scenario, of course, refers to the bandwidth and not the technology that provides it. When asked whether the company might be interested in releasing any fraction of its two million lines of proprietary code into the public domain as open source, Furrier demurs.

“Haven’t really thought of that,” he said. “Right now we want to deliver faster, cheaper broadband.”

EtherLinx Reaches Out

All of this future progress, though, is clearly going to be dependent on key business relationships at every stage of the products -and-provisioning game. Are those relationships in place — or will they be soon? In other words, is EtherLinx in fact currently negotiating with major carriers, WLAN solution providers, or is all the media attention just a tempest in a teapot?