Cohda Wireless
Cohda was spun out of the Institute for Telecommunications Research (ITR) at the University of South Australia in Adelaide in 2002. Last fall, the company took on an American CEO, Martin Suter, and opened its headquarters in Orlando, Florida. It has been adapting technology originally developed at the Institute for proprietary networks and applying it to Wi-Fi. The company claims the technology – enhancements to the receiver section in Wi-Fi radios – solves propagation problems that arise when Wi-Fi is used in an outdoor, high-speed mobile environment.
The strategy is to license its technology in the short term to outdoor Wi-Fi equipment vendors to help them increase the range of their access points. In the longer term, Cohda hopes to sell the technology to chip vendors who will incorporate it into both access points and client devices. It has done neither yet.
“There are advantages to using Cohda devices in infrastructure equipment,” Suter says. “We have a value proposition and we’re out talking to the systems guys. However to achieve maximum benefit, you need to make changes at both ends of the link.”
The technology is, at the very least, interesting. The marketing campaign, however, is bizarre, in that it seems designed to insult some of the very constituencies the company hopes to convince of the necessity and merits of its technology – Wi-Fi system vendors and municipalities.
In the press release announcing the Dirty Little Secret site, Suter is quoted as saying, “The ubiquity and ‘low cost’ of Wi-Fi has fooled people into believing it is an attractive option to meet the demand for outdoor, metro-area broadband wireless networks. However, 802.11 was not designed for mobility or for outdoor deployments, especially in cities where harsh multipath from clutter, such as buildings, cars, and trees, is a major issue. Overcoming these issues requires saturation of an area with 802.11 devices – but this dramatically increases costs and obliterates return on investment. 802.11 can be made more economically viable, but customers will need to push vendors to make it happen.”
The site and the release reference problems and uncertainties with high-profile muni-Wi-Fi projects in San Francisco; Chaska, Minnesota; and St. Cloud, Florida. Business plans in those cities have indeed been thrown into disarray at times, or at least cast into doubt, because of the discovery that proposed node densities cannot deliver the level of service municipalities want.
The implication in the language used at the Cohda site is that the industry has deliberately misled municipalities or obfuscated the issues. Isn’t there a chance this imputation could really tick off prospective customers, we wondered?
“Sometimes you need to poke somebody in the eye to get their attention,” Suter says. “Yes, I will run the risk of pissing several of them off to get their attention, but when we begin the discussion, it’s impossible for them to dispute the problems.”
Veteran telecom consultant Phil Belanger, an expert in muni-Wi-Fi and managing director of Novarum, a firm he formed recently with partner Ken Biba, is probably not alone in finding the Cohda marketing slightly off-putting.
“I don’t like this whole concept of the ‘dirty little secret,’” Belanger says. “I don’t think there is a dirty little secret. [Cohda] is saying in effect that the industry is conspiring to fool cities [into using Wi-Fi]. There is no conspiracy.”
He doesn’t dispute Cohda’s assertion that Wi-Fi, based purely on its technical merits, is far from the best candidate for providing citywide mobile broadband service. “But the industry is not conspiring to hide that fact,” he insists. Wi-Fi is being used in spite of its acknowledged shortcomings in outdoor and mobile applications because it is ubiquitous and therefore cheap, so it does, or at least can, support viable business cases for citywide broadband services.
Belanger also contends that Cohda is making misleading use of some of its examples of muni-Wi-Fi gone wrong. In one case, he says, the network was apparently badly designed from the get-go. In another, the problem was mis-management of expectations. “Every grandma in the city thinks she can get broadband access anywhere in the house if her address is St. Cloud,” he says. “So now the city is throwing money at infrastructure to make that true. Those examples don’t really apply.”
The apparently wrong-footed marketing approach is all the more difficult to understand given that Suter is an articulate and sober-sounding executive who has been working on the problem of mobile broadband since 1999 – initially with Nortel Networks, then MeshNetworks (since purchased by Motorola) and finally with Motorola. He makes a plausible case for why and how “plain vanilla” Wi-Fi, as he calls it, is failing in citywide broadband initiatives, especially in mobile applications. He also sounds plausible when talking about why system vendors and chip makers aren’t doing anything about it.
The technical problem, in simple terms, arises from the fact that a wireless signal arrives at a receiver by multiple paths, including as a result of the transmitted energy reflecting off physical obstacles. These multipath signals arrive later than the first received signal. In the case of an indoor setting, where the signal is bouncing off nearby walls and furniture, they arrive only a few tenths of a millisecond after the main signal. Outdoors, the delay may be significantly longer, and if the sender is mobile, that further complicates the multipath problem.
To help simplify the process of tuning a signal, Wi-Fi radios set an arbitrary limit on how long after the main signal arrives the radio will continue to attempt to resolve multipath energy as coherent signal – a mere .8 milliseconds. After that, it treats multipath as noise. The further the sender is from the receiver, the longer the multipath delays, so when Wi-Fi goes outdoors, radios in effect throw more energy away. And this is a serious limiter of both signal robustness and – more importantly, Cohda would argue – range.
“There is a misconception,” Suter says, “that [the multipath problem] is a power issue. But there is available power out there. It’s really a radio coherence issue, and that is as a result of the intolerance of conventional radios to the delay spread.”
Cohda’s technology essentially disregards the delay spread stipulated in the Wi-Fi standard and uses all of the received energy generated by the sender to resolve a coherent signal, including multipath energy a conventional radio discards as noise. The trick was to do this without introducing additional signal latency, which Cohda claims to have achieved. Suter naturally won’t give a lot of detail about how, other than to say the technology uses an “iterative processing paradigm.”
“What it means in effect is that you can create a very robust link,” he says. “You take energy that normally would be incoherent and make it coherent.” And that translates to significantly increased range, so municipal broadband services could make do with fewer access points to cover a given area.
Belanger, however, questions the fundamental logic of increasing the size of cells in a muni-Wi-Fi network. Each access point can only handle so many simultaneous users, so if the network is successful, it will eventually need more APs to handle all the users demanding access. The node density most muni-Wi-Fi networks are using today, about 25 to 30 APs per square mile, is probably about right, he believes. “What you need is more robust signals in those areas where you have lots of traffic. You don’t want bigger service areas because of the capacity issue.”
Cohda, meanwhile, has built a demonstration network around its research center in Australia using prototype equipment. Suter says the company also proved the technology’s effectiveness in a demonstration at a recent conference of the Association of Public-Safety Communications Officials (APCO).
Then why haven’t vendors rushed to solve the multipath problem in Wi-Fi, if it’s so readily solvable?
Suter says it’s because the logical way to solve it is at the chip level and the chip vendors don’t see an economic incentive to do it, at least not in the short term. This is because the problem only affects – and the solution only helps – users of outdoor, mobile networks. That, despite the recent rise of muni-Wi-Fi, is still a relatively tiny part of the market. And as one chip maker recently explained to Suter, margins are wafer thin.
“[The semiconductor industry] is completely driven by cost and volume,” he says. “This guy told me his company could maybe support pennies in incremental cost [per chip to implement something like Cohda’s technology] but not dimes or quarters, because they’re bidding on multi-million dollar contracts with companies like Cisco, and anything that increases their cost that isn’t of interest to the volume market doesn’t make sense.”
Suter believes that when shipments of Wi-Fi-enabled cellular handsets reach volume levels in two or three years, that will be a tipping point at which the market dynamics his company is struggling with today will change. “The desire of [owners] to use those devices wherever they are will result in an exponential growth in the outdoor Wi-Fi market,” he predicts.
At that point, enhancing chipsets with technology like Cohda’s will become a more economically viable proposition for chip makers. Suter’s hope is to find a chip maker with a similar vision of the future who recognizes that using the Cohda technology will provide a competitive advantage when the outdoor Wi-Fi market really takes off.
The company might even succeed, if it can just keep its foot out of its mouth.
