A giant neon cowboy beams down at us. It's Vegas Vic, the smiling, smoking, mechanical icon of downtown Las Vegas. I'm sitting in the back seat of an SUV driven by Mitchell Gonzalez, president and founder of Cheetah Wireless Technologies Inc., a small start-up located in a cookie-cutter business park directly opposite Las Vegas's busy McCarran International Airport. I'm surfing the Web at healthy broadband speeds—1.5 megabits per second or so—as we drive around a 5-square-kilometer patch of the city where Cheetah is adding the finishing touches to a pilot mobile broadband wireless network.
I'm here because this little district of casinos, hotels, wedding chapels, and souvenir stores could be the ground from which the seeds of a telecommunications revolution will grow. Cheetah's mobile network is one of the first municipal installations to use mesh wireless technology, which will allow users to access the Internet anywhere within the coverage area—even if they're driving at 100 kilometers per hour. If the technology is adopted by the City of Las Vegas and other municipalities beyond, it will herald the arrival of a major player in mobile broadband, leapfrogging cellular technologies and next-generation WiMax.
Municipal broadband wireless networking is a market currently worth almost US $500 million in the United States and one that could grow to over $2 billion by 2008, according to Input, a Reston, Va.-based analysis firm. And if mesh operators can establish themselves in towns and cities across the nation, selling services to commercial users in addition to municipal ones could swell their revenues even more. A study by two other firms, BWCS, in Ledbury, England, and Senza Fili Consulting LLC, in Sammamish, Wash., estimates the commercial U.S. wireless broadband market will grow to $3.7 billion by 2009.
Mesh networks promise several key advantages over traditional wireless solutions, such as Wi-Fi or cellphones. Benefits include higher speeds, less susceptibility to radio interference, and greater resistance to network congestion. These networks also offer better coverage, the ability to prioritize different types of users, geolocation capabilities, tighter security, faster deployment, and a degree of immunity to catastrophic network failures. And, perhaps best of all for cash-strapped local governments, mesh network vendors are willing to be creative about financing.
Cities and towns possess something that can be even more valuable than cash to vendors of an upstart technology looking to take on the cellphone companies' current lock on the mobile data world: poles. Streetlights, traffic signals, and road signs—they're all attached to city-owned poles perfectly positioned for deploying the backbone of a wireless network, and as an added bonus, a lot of them are already wired for electric power.
In contrast, cellphone stations have to arrange for their own power and be placed on costly towers or attached to pre-existing buildings, often entailing lengthy and expensive negotiations, especially when there's local opposition.
So the wireless industry and municipal governments alike are keeping a keen eye on a dozen or so pilot programs that are carrying out mesh network demonstrations around the United States. With major cities like Philadelphia and New York in the planning stages of what will be massive, citywide networks, the stakes are high.
But this was far from the mind of Jorge Cervantes, assistant traffic engineer for Las Vegas, when, in 2003, he started casting about for a way to prevent drivers from abusing the system used to change traffic lights to green as emergency vehicles approach.
When black-market devices started appearing that let any driver trip the lights, the Las Vegas Traffic Engineering Department decided to acquire a new system that used coded signals so that only authorized vehicles could control the lights. But with 500 intersections and a constantly changing fleet of emergency vehicles, city officials rejected the idea of keeping each intersection's traffic-light system manually updated with the current database of authorized vehicles.
"We were looking for a way to communicate to all the signals, and we got introduced to Cheetah," explains Cervantes. "As we explored what they had to offer, we saw additional potential uses for mesh technology."
As an example, he cited the city traffic department's network of cameras to monitor traffic flow around the city. The mesh network could be used to bring images from the cameras back to the traffic management center. It could also help municipal work crews, equipped with mesh-enabled laptops out in the field, to download instructions and schematics from city databases.
Before long, interest began to spread to other city departments. For emergency services, for instance, Cheetah recently staged a demonstration of how responders could use the network to communicate during an emergency that included videoconferencing in the field. Cervantes has lent the fire department some mesh cards so it can start testing the pilot network.
The mesh network also offers an opportunity for Las Vegas to solve its interoperability problem. What visitors think of as a single glitzy metropolis is actually a patchwork of municipalities and areas that fall under the control of the surrounding county. Even the Las Vegas Strip, the city's signature avenue of outlandish hotels and casinos, is nearly all part of an unincorporated township called Paradise.
As a result, when major problems arise, it's difficult to coordinate the different agencies involved, such as the fire and police departments. "We've got four or five different cities, and it seems that each city has a different communications system," says Cervantes. "The mesh network could be a way to bridge that interoperability gap" if each city joined the mesh network.
Cervantes and Cheetah worked out a deal to start building a proof-of-concept network last summer. The Las Vegas traffic department agreed to provide most of the manpower and the lampposts needed to deploy the network hardware and pay Cheetah $25 000 to defray its costs. Cheetah provided all the equipment but retains ownership of it, pending the city's decision on installing a full-scale network.
Cheetah refuses to disclose the per-unit cost of the equipment but said the Las Vegas pilot network cost about $175 000—or roughly $80 000 per 2.6 km2. Of that, $30 000 pays for things like the servers Cheetah uses to manage the network and the cost of connecting it to the Internet, but the bulk, $50 000, is the cost of the mesh hardware itself.
Cheetah's hardware comes from Maitland, Fla.-based MeshNetworks Inc. [see "10 Tech Companies for the Next 10 Years," IEEE Spectrum, November 2003]. As this article went to press, communications technology giant Motorola Inc. announced it had signed an agreement to acquire MeshNetworks, with the stated intention of introducing mesh technology across all of Motorola's business units, from home entertainment to cellphones. While Motorola has its own sales force, licensed resellers such as Cheetah have been grandfathered into the agreement, says Rick Rotondo, MeshNetwork's vice president of marketing. He expects that many of MeshNetwork's current resellers will be used on a contract basis by Motorola to deploy and operate future mesh systems, because of their experience and wealth of regional contacts.
Cheetah also sells two other types of mesh networks based on hardware from Tropos Networks Inc., in Sunnyvale, Calif., and BelAir Networks Inc, in Kanata, Ont., Canada. The Tropos and BelAir mesh systems are built on top of the IEEE 802.11 Wi-Fi protocol, while MeshNetworks uses a proprietary technology originally developed for military applications.
Compared with the Wi-Fi-based mesh systems, the military-derived technology offers improved resistance to radio interference, better security, built-in geolocation, and the ability to offer different levels of service to different types of users. It also allows users to move about seamlessly within the entire coverage area, instead of having to re-establish a connection as they move from one Wi-Fi hotspot to another.
Within the Las Vegas MeshNetworks system, the average transmission speed ranges from 500 kilobits per second to 1.5 Mb/s, with bursts of up to 6 Mb/s possible. The backbone of the network is made up of 33 or so shoe-box-size gray boxes attached to traffic-light poles high above the streets, known as wireless routers [see photo, " "]. During the pilot deployment, city workers with a bucket truck managed in one 8-hour shift to put up 18 routers, says Cheetah's Gonzalez. Routers are simply bolted to lampposts and plugged into the photocell power adapter that sits atop most streetlights. The routers are then automatically assimilated into the mesh network. The most time-consuming element of installing or removing one is the time required to put traffic cones around the work area.
Automatic assimilation is the essence of what makes a mesh network tick and what makes it different from a standard Wi-Fi or cellular network, which follows a hub-and-spoke model. The hub-and-spoke model has a sharp distinction between hardware such as cellphones (spokes) and cellphone towers (hubs). In this situation, if two cellphone users want to talk to each other, their conversation must be relayed through the tower, even if they're sitting in adjoining rooms and the tower is a kilometer away.
Problems arise with the hub-and-spoke model as more users try to access the same hub. For example, if too many Wi-Fi users try to connect to the same wireless access point, network congestion can slow traffic. Eventually a point will be reached where no more users will be able to access the network at all. By the same token, if a conventional cellphone tower stops working because of a power failure or terrorist attack, users will be left unable to communicate with one another or the rest of the world.
But once brought into range of a mesh network (and authenticated as having permission to be part of the network), any piece of mesh hardware—such as a wireless router, an interface card for a laptop, or a mesh-enabled VoIP-based (for voice over Internet Protocol) radio—is assimilated into the network and starts acting as a relay. Every piece of equipment maintains a constantly updated so-called routing table that describes the best way of sending data from one point on the network to another.
As far as a piece of traffic on the network is concerned, there's no distinction between a router and an interface card: each can receive and forward transmissions from its neighbors on a peer-to-peer basis. Instead of looking like a hub-and-spoke setup, the network looks like a mesh—hence the name [see diagram, " "]. If the structure looks familiar, it should—the Internet has a similar, decentralized, mesh structure.
Though Wi-Fi devices and cellphones can be set up to communicate directly with each other instead of through a hub, there are several serious limitations. Without a mesh's ability to automatically route traffic, each device can communicate only with other devices within its immediate broadcast range.
To see why this is a problem, imagine that Alice, Bob, and Charlie are all trying to send each other Instant Messages wirelessly with Wi-Fi-enabled laptops configured to talk to each other directly. They're spread out so that while Alice and Charlie are both inside Bob's broadcast range, Alice and Charlie are out of range of each other. Alice and Bob can chat to each other, and Charlie and Bob can chat away, too, but Alice and Charlie can't communicate directly. And if Alice's computer has Internet access, say, through a separate Ethernet connection, Bob (let alone Charlie) can't reach the outside world through her connection.
But if Alice, Bob, and Charlie were using mesh technology, Charlie's laptop would know that in order to reach the Internet, data would have to go to Bob's laptop, which would forward the data to Alice's machine, which in turn would send Charlie's data out to the Internet. And as more users joined the mesh network and started routing traffic within their broadcast range, coverage would automatically improve, thanks to these additional relays.
This characteristic is enormously important to municipal officials worried about a terrorist attack. The reason is that, in a disaster, emergency workers equipped with mesh-enabled hardware would be able to communicate with each other—even if the local communications infrastructure were destroyed, as happened with the World Trade Center collapse on 9/11. There would be no need for the radio repeaters currently used in tall buildings to keep firefighters in touch with command posts. Apart from each firefighter's radio acting as a relay station, additional mesh-based relays could be dropped every few floors, possibly with sensors monitoring conditions, such as temperature and carbon dioxide concentrations. And connectivity to the wider world, such as the local emergency headquarters, could be restored by, say, parking a few mesh-equipped police cars along a route between the scene of destruction and the rest of the network.
"If you look at events like the World Trade Center, you certainly would be able to recover from things like that automatically" using mesh technology, says Scott Midkiff, an electrical engineering professor at Virginia Tech in Blacksburg. In fact, Midkiff notes that the World Trade Center site provided an early test of mesh technology's utility in a disaster. At the height of the dot-com boom, a short-lived start-up called Ricochet Networks had deployed mesh networking technology throughout Manhattan and other metropolitan areas. While Ricochet succumbed to the bust two months before 9/11, its infrastructure was still in place around Ground Zero and was reactivated after the attacks for the benefit of recovery personnel until normal communications could be restored [see "What Went Wrong at Ricochet," Spectrum, March 2002].
Can Cheetah succeed where Ricochet failed? Certainly, a lot will hinge on the technical details. In a mesh network, the role of Alice's Internet-connected computer is played by an Intelligent Access Point (IAP). This is where the wired world meets the wireless, and it is usually the most difficult part of the network to install, because a high-speed connection must be brought to the IAP.
In Las Vegas's case, this is a T-1 line for each of three IAPs that links them to Cheetah's servers at a secure facility. An IAP takes about a day to install, assuming the T-1 is already in place. Besides providing Internet connectivity, the servers also verify whether a piece of hardware, such as a laptop card or wireless router, should be assimilated into the network, keeping interlopers out. The central servers can also set priority levels for different users—emergency first responders, such as the fire department, would get the highest priority, while a road-repair crew would get a lower priority.
One IAP is needed for every 2.6 km2 or so, and one wireless router is needed about every 0.62 km2 to seed the network [see " "]. Local conditions can dictate the need for more routers. For example, the city of Medford, Ore., is also currently evaluating a municipal mesh network. But Medford, with a population of 70 000, is in a valley wooded with pine trees, and engineers quickly discovered that more routers were needed than originally thought. At the frequencies used by MeshNetwork's equipment, "pine needles absorb a high percentage of the signal," Doug Townsend, Medford's technology services director, tells Spectrum.
In Las Vegas, the problem isn't so much trees springing up to block radio waves as it is buildings. Las Vegas is in the midst of a construction boom—during the 1990s, it was the fastest-growing city in the United States—and Cheetah has plans to deploy five more routers in its test area to keep up.
Another problem separates Las Vegas from Medford—chronic radio interference. "We joke that it's because we're the closest city to Area 51—the aliens must be creating all the interference," quips Guchi Sacramento, a network engineer for Cheetah. But "actually, the airport has a lot to do with it."
Yet more interference comes from the heavy use of Wi-Fi by businesses in the area. Relying on the same unlicensed portion of the spectrum between 2.4000 and 2.4835 GHz that Wi-Fi uses, the mesh network has to share the airwaves with a lot of other users.
MeshNetwork's hardware gets around the interference problem by using a spread-spectrum approach that sends data over multiple channels. This technique also makes it a lot harder to eavesdrop on network traffic. Without an authorized MeshNetworks card, it would be difficult to reassemble the radio signals into the original data stream, a problem further compounded by the fact that different parts of a message can flow through different computers as they make their way through the mesh network.
Because the physical location of the IAPs and wireless routers can be precisely determined, a roaming user's interface card can determine its position by noting which routers are within broadcast range to an accuracy of about 10 meters. That's a handy bonus for municipalities that have invested heavily in Geographic Information System (GIS) technology to keep track of their infrastructure and monitor local conditions.
But smart technology won't be enough to make mesh networks mainstream. For wide-scale adoption by municipal governments, they have to be affordable in a time of tight budgets. New homeland security funding in the wake of 9/11 is helping to some extent. Medford, which bought its mesh network outright from another reseller of MeshNetwork hardware, Viasys Corp., in Lakeland, Fla., got some $500 000 of the $700 000 price tag for its pilot program from the federal government.
Many mesh vendors are helping to bridge the financing gap with creative public-private partnerships, and Cheetah is no exception. Rhonda McGarran, Cheetah's vice president of marketing, explained its proposition: "A mesh network for a small city might cost $20 million. Well, because that city doesn't have $20 million and it's not going to be able to raise it or doesn't want to wait until it can, there's another way that we can do this."
Vega's metro mesh network
Goal: To deploy a pilot program for a citywide broadband mobile wireless technology
Why it's a winner: The robust and easy-to-deploy mesh technology allows a city to serve a wide range of needs, from emergency responder communications to traffic-light management
Organizations: Cheetah Wireless Technologies, MeshNetworks, and the City of Las Vegas Traffic Engineering Department
Center of activity: Downtown Las Vegas
Number of people on the project: A few dozen
Budget: Approximately US $175 000 for the pilot deployment, $6 million for a full-scale city deployment
Cheetah's typical proposition is for the municipality to give the company the right to attach its equipment to city-owned light poles and establish Wi-Fi hotspots around the city. These hotspots act as bridges that allow users with regular Wi-Fi cards to access the mesh network, and they let Cheetah tap into the consumer market—a typical Wi-Fi laptop card, of which millions have already been sold, costs about $80, while a mesh card is more like $800.
With the city's help in promoting the service, Cheetah would charge users between $20 and $40 per month to access fixed Wi-Fi hotspots. Commercial users who want to access the mesh network directly so as to be able to roam the entire coverage area seamlessly will be charged about $60 to $80 per month. Revenues are shared between the city and Cheetah, with a portion of the city's cut set aside to pay for the municipality's use of the mesh network. "With enough subscribers the city could use the network at no charge and could even start getting revenue from it," says Gonzalez.
But all the clever financing in the world means nothing if cities aren't happy with the results of their pilot programs. For Las Vegas, Cervantes stresses that it's too early to tell, but he's clearly enthusiastic about the technology and notes happily, "We're pleasantly surprised that the pilot deployment was very easy to do."
A decision on whether or not to roll out mesh to the entire City of Las Vegas—some 150 km2—is expected in the next month or so. And, with a gleam in his eye, Cheetah's Gonzalez hopes to extend mesh coverage to the entire surrounding county—over 20 000 km2—in the next few years.
In Medford, Townsend is "very pleased" with how the project is shaking out and notes that some of the test users in the fire, police, and public works departments are "ecstatic" about the system. The city has already secured another homeland security grant for the next phase of development and started boasting about its network in advertisements designed to attract businesses to the area.
Interest among other cities is clearly high. Townsend has had visitors from local governments as far away as Redding, Calif., and Hattiesburg, Miss. Cervantes also reports fielding queries from other municipalities.
It's clear that unless the cellphone companies can produce something spectacular in the next 12 to 18 months, or there's an equally spectacular mesh network failure, municipal mesh is the future—at least in rural and suburban areas where the communications infrastructure isn't as well developed as in larger metropolitan areas. In these larger cities, cellphone companies are already offering mobile broadband connectivity, albeit at lower data rates than mesh can provide and without many of mesh's disaster-resistant features.
Still, mesh is proving competitive in these big cities as well, with Philadelphia announcing last May that it was expanding a trial of a free public Wi-Fi-based mesh network to cover famous city landmarks, and Motorola's acquisition of the technology is a big boost. In New York City, officials are currently examining proposals for a public-safety wireless network to blanket the Big Apple. Should they decide to build it with mesh, it will mark the technology's coming of age.