Wireless mesh emerges as a viable wireless broadband alternative for metropolitan areas and campus environments

Mesh takes a higher profile: wireless mesh emerges as a viable wireless broadband alternative for metropolitan areas and campus environments

John C. Tanner

Wi-Fi’s strength as a broadband access technology has always been offset by its poor range. For all its early plaudits as a 3G killer, the truth was that Wi-Fi simply didn’t have the signal range to compete with W-CDMA and 1x EV-DO. Even as a localized hotspot service, Wi-Fi wasn’t as inexpensive a service as it looked. Access points were cheap, but the costs of backhaul links and truck rolls for each AP, as well as the back-office requirements, were not.

A solution in the works for several years now has been wireless mesh, which essentially proposes turning APs into router nodes with peer radio devices that can automatically self-configure and communicate with each other. They can also self-learn changes in the network (say, when a new node is added or removed) and adjust transmission paths accordingly for optimal throughput.

More importantly, says Billy Basu, Nortel Networks’ senior manager of telecom and enterprise networks, there’s no need to provide individual backhaul connections for each and every AP as required by regular Wi-Fi Aps. You can connect a few nodes to the PSTN and allow the others to provide their own backhaul links to each other.

“You don’t have to hardwire every access point,” he says. “Instead, the access points connect with each other, so you get any-to-any connectivity that can extend hotspot coverage over a much larger area. It’s a very cheap way of deploying broadband, especially in towns and cities where broadband rollouts are limited or even non-existent.”

Despite some initial interest in mesh, the technology slipped under the radar for a while as everyone’s attention turned to technologies like WiMAX, which has been rather simplistically viewed, more or less, as Wi-Fi with much better coverage.

DORMANT NO MORE

But wireless mesh hasn’t been dormant amidst the WiMAX hype. Start-ups like Tropos Networks and MeshNetworks have been deploying wireless mesh in municipalities across North America. The highly publicized legal fracas in Philadelphia over plans by the city government to deploy a public Wi-Fi network using mesh has called new attention to the possibilities of mesh as a wireless broadband solution.

So has the fact that mesh is no longer being driven by start-ups. Big names like Motorola and Nortel Networks are pushing the mesh message.

Indeed, Motorola bought MeshNetworks last year and partners with Tropos to provide mesh solutions as part of its Canopy wireless broadband technology.

Nortel, meanwhile, has been raising mesh’s profile in Asia, scoring contract wins in Taiwan and Australia. The city governments of Taipei and Kaohsiung have both deployed wireless mesh across their metro areas. Wireless mesh is also appearing on campuses.

Such examples highlight the fact that wireless mesh is virtually a DIY approach to wireless broadband that has had particular appeal to city governments, and enterprises and universities with a campus environment and an interest in extending their WLANs outdoors. This doesn’t mean service providers aren’t interested, says Joshua Chao, Asia Pacific sales vice president for Proxim.

“Service providers are interested in providing for mesh projects that have been initiated by governments … because the governments might want to build these networks, but they don’t necessarily want to manage them,” Chao says. “Cities can provide the locations to allow the network to be built, then the actual network can be operated by a service provider.”

How far wireless mesh goes remains to be seen, with all eyes still on WiMAX and similar technologies. However, mesh has several things going for it over WiMAX. One is the growing emphasis within the wireless industry on co-existence of wireless broadband technologies. The other is that WiMAX as a mobile-enabled solution is still a few years off. Mesh is available today.

SCALABILITY CONCERNS

That’s not to say wireless mesh isn’t without its drawbacks. In fact, mesh has had its own deployment issues. Some have been overcome, like the baggage of security concerns over Wi-Fi in general, says Basu of Nortel.

“Wi-Fi had some security issues previously, but with the WPA standard now available, most of the problems have been eliminated,” he says. “Nortel’s wireless mesh equipment uses IPsec for data, and we’ll also have 802.11i later this year, so the security issues have largely been dealt with.”

Another issue for wireless mesh has been scalability, though it’s important to define what one means by scalability in a mesh network.

For example, Lonnie McAlister, product line manager of wireless networking at Intel Communications Group, points out that wireless mesh is scalable in terms of coverage, but not data rates.

“It’s inherent in the technology,” he explains. “The further you get from the access point, the lower the data rate. 802.11 isn’t scalable in terms of the number of users you can fit on one access point. Usually it tops out at 25-30 users.”

Lynn Lucas, vice president of product marketing at Proxim, says that a wireless mesh network’s scalability depends on the number of radios in the AP.

“For example, single radio solutions aren’t very scalable,” she says. “Moving away from the access point has a dramatic impact on throughput. You’re basically splitting the radio between client traffic and backhaul traffic. If you have multiple radios for client traffic and backhaul, you’ll get broader coverage.”

That said, Lucas adds that Proxim’s mesh customers are less worried about scalability and more about reliability.

“Our customers are approaching mesh more like cellular operators do. They’re not looking at the scalability so much as they are at the impact on service if a base station goes down and how many users would be affected.”

MANAGING INTERFERENCE

One potential worry for wireless mesh is radio interference, due to the fact that 802.11b and 11g operate in unlicensed spectrum, says Basu.

“If you build a giant Wi-Fi grid, especially using unlicensed spectrum, you’re going to have interference issues,” he says, adding that this is one reason why service providers haven’t quite embraced mesh. “That’s why we view it more as a campus application. Service providers would have to be careful how they develop this.”

On the other hand, Basu notes that in markets where a spectrum license is required, interference will be much less of an issue for license holders.

Intel’s McAlister agrees.

“Unlicensed spectrum is good for campuses, but carriers want to use licensed spectrum that they have more control over.”

McAlister says that sources of potential interference could include Wi-Fi-based home networks, which are expected to become even more commonplace over the next few years. However, Lucas doesn’t see this as a problem.

“On a practical level we haven’t seen it for a couple of reasons. One is that Wi-Fi in the home will be attenuated severely once it goes outside, and vice versa,” Lucas says.

The second reason, she adds, is that more mesh APs are now using 802.11a as the backhaul frequency while running client access traffic on 11b and 11g, which allows for load balancing that keeps interference to a minimum.

The issue of backhaul spectrum is where WiMAX is expected to eventually come into play, not as backhaul between nodes, but as a backhaul link from the mesh network to the PSTN, especially for mesh deployments in more remote locations, says McAlister.

That’s the 802.16d flavor of WiMAX, which only supports fixed wireless links. 802.16d fits well into the view of WiMAX as a complementary technology to mesh. The trick, says Lucas, is what happens when 802.16e, which will support mobile WiMAX users, is finalized.

“802.16e is still several years out and hasn’t even been ratified yet,” she says. “But looking at 16e vs wireless mesh, it’s still unclear what will happen. I think it will depend on the client side and how that develops. Clients like laptops, PDAs and eventually even phones will have to support 16e, and we haven’t really seen a roadmap for that yet or whether it’ll be embedded like Wi-Fi today or not.”

Which brings up one of the key advantages wireless mesh has over WiMAX: It’s here now.

“Standard and interoperable WiMAX isn’t here yet commercially, and the standard for supporting mobile users isn’t finished,” says Basu. “And until it is, and until the antennas and base stations are available at a reduced cost, we won’t see massive rollouts anytime soon.”

McAlister agrees. “Much of the interest in mesh today is due in large part to WiMAX equipment availability, or lack thereof.”

PARTIALLY PROPRIETARY

Ironically, wireless mesh itself isn’t standardized either. Not exactly, anyway. The radio interfaces for 802.11a/b/g are all standards, but the actual technology that enables the access points to mesh is not, says McAlister.

“The value-add in mesh is getting access points to recognize each other, and to reroute intelligently if an access point goes down. This is proprietary technology in nature,” he says.

But, he adds, that’s just between the access points. “To an end-user, it’s invisible. Their 802.11b/g clients will still work on a wireless mesh network, so it doesn’t force them to buy into some proprietary client hardware.” As for the infrastructure side, Lucas acknowledges that the actual meshing aspect is proprietary, but says it’s not an issue with customers.

“It’s only an issue on the infrastructure side, and we don’t see many enterprise customers mixing access points from different vendors. And this has no impact at all on the client side, so it’s not really a deterrent.”

Depending on the customer, proprietary mesh can even be a bonus. Dennis Stipati, director of international markets for Motorola’s Canopy Wireless Broadband group, points out that while mesh technology from its new subsidiary MeshNetworks is proprietary, this is an asset since its solution is targeted at public safety customers.

“The police like that it’s proprietary because they’re not crazy about unauthorized people being able to access the network with a laptop or PDA,” he says.

KILLER APP

One of the more interesting, and potentially controversial, prospects of wireless mesh is the emergence of Wi-Fi handsets and the ability of Wi-Fi to support voice. The wireless mesh network at Edith Cowan University in Australia, for example, offers voice services to students, who can use Wi-Fi/cellular handsets to make intra-network Wi-Fi calls on campus and roam to cellular for off-campus calls.

“The value proposition for voice from the university’s point of view is that it turns a cost center into a revenue center,” says Basu of Nortel, as it not only saves money on the cost of networking the campus but also deploys it as a subscription-based service.

Basu has no doubts that voice will be a killer app for mesh. “We’re already seeing dual-mode [Wi-Fi/cellular] handsets, and the battery power has improved a lot since the first ones came out. Because of VoIP and SIP, we’ll start to see more voice running on Wi-Fi, especially in campus deployments.”

It’s this sort of thing that has telcos worried in some markets where broadband operators are already competing in the local telephony space with VoIP. However, Basu says the threat is overblown, as VoWLAN is more suitable for campus-based mesh networks than, say, public metropolitan systems.

“Voice over Wi-Fi is a killer app, but it’s constrained by engineering rules,” he says. “It’s easier to manage on a subscription-controlled campus network. It’s certainly never going to replace cellular.”

Voice or no voice, the prospects of wireless mesh are promising, if unclear.

“There’s more interest now that you can actually see the technology deployed and working,” Stipati says. “That’s how it usually goes. No one wants to be the first to try it, but once they see someone else try it, then they go for it.”

Pros and cons of mesh:

Pro:

* Cheaper way to deploy broadband where existing rollouts are limited

* Planned rollouts in U.S. and Asia highlight technology

* Large, band-name vendors now embracing mesh

* Technology available now

* Service providers needed to manage the mesh networks

Cons:

* Poor signal range compared to W-CDMA and EV-DO

* Relatively expensive backhaul links and back-office requirements

* Wi-MAX standards’ agreement could pose challenge

* Scalability of mesh is questionable

* Radio interference is a concern

COPYRIGHT 2005 Questex Media Group, Inc.

COPYRIGHT 2005 Gale Group