|
|
Wireless mesh networks are proliferating rapidly, in vertical applications, temporary deployments for events like NASCAR races and municipal networks. While the architecture can be used for everything from setting up ad-hoc peer-to-peer communications on a military battlefield to providing in-building broadband coverage for hard-to-wire locations, there have been louder rumblings lately about using mesh to provide metro broadband access for cities. With a lack of standardization, multiple radio and frequency approaches, and a critical need for market education all hanging over their heads, purveyors of mesh architecture must frame the value proposition without getting bogged down in mesh’s variations.
The mesh concept is straightforward: Instead of deploying each access point with a wired connection that can handle a given number of simultaneous users as in a traditional WLAN, entities using mesh architectures take advantage of an access point or wireless radio with router functionality. This combination router/access point is known as a node, which locates and connects to every other node in a network to transmit packets between them on a multipoint basis. It requires only a few wired backhaul connections to provide shared bandwidth to all nodes in the network, which in theory means it can cover greater distances at a lower cost than a traditional WLAN, particularly in dense cell configurations. Radio devices may have access to one radio channel, which means they cannot send and receive packets at the same time, or could have dual or multiple radios.
Click to Enlarge |
“Reactions to the mesh concept have evolved in people’s minds from incredulousness, to thinking it might work, to being curious about successful installs,” says Bert Williams, vice president of marketing at Tropos Networks Inc. The Tropos MetroMesh architecture provides metro-scale broadband wireless coverage using single-radio Wi-Fi access points, centralized element management and a purpose-built hardware platform. “While the market is in its infancy, we see metro deployment for service providers and municipalities as our main space going forward.” Tropos is working with Intel Corp. on its Digital Communities initiative, designed to help municipal governments use a wireless infrastructure to expand and improve services for businesses and citizens by designing, developing and deploying citywide wireless networks. Corpus Christi, Texas, is a pilot community in the program, using Wi-Fi mesh for automated gas and water meter reading, giving public safety workers mobile access to their desktops and other productivity applications.
Click to Enlarge |
Nonetheless, the market is filled with a jumble of customized gear and configuration concepts. For instance, radios provide connections to laptops or other wireless devices and provide a path for client packets to and from the wired network “root.” In large-scale deployments, there may be more hops for the data across the mesh until the backhaul network is reached. In single-radio configurations, a node first receives the data and then retransmits it, resulting in a packet loss for each hop — the more hops, the more degraded the signal becomes. In a multiradio scenario, a node can send and receive at the same time, minimizing packet loss and providing real-time back-and-forth required for VoIP and IP video.
“The technology is still young, so there is of course room for improvement,” says Williams. “We are working on increasing capacity to provide more bandwidth per user, and we’re working on establishing QoS for voice handling ability.” He says Tropos has optimized its proprietary software to provide solid end-to-end throughput in a single-radio configuration. “We get the same performance at a fraction of the operational and installation expense of a multiradio approach,” he says.
Other manufacturers prefer a multiradio approach, where one channel can receive and another can transmit, for real-time communications.
“We take a normal AP footprint and add a lot of bandwidth and application support,” says Cyrus Irani, co-founder and vice president of marketing and strategy at Strix Systems Inc., whose Access/One Network family of products provides a distributed networking infrastructure and high-capacity nodes with one to six radios. It released its six-radio product for outdoor hot zone deployments in April.
“Our nodes ask, ‘Who am I and where am I?’ then self-configures with assistance from a network server. They create a mesh backbone for the data path, then self-tune and load-balance. The end result is that we can move a lot of data quickly in a multipoint-to-multipoint fashion, with very little interference.”
Click to Enlarge |
Players acknowledge that even with a multiradio approach, some kinks need to be worked out. “The market is very early,” says Irani. “We still need education in the market on what is involved. You need 60 to 70 nodes and a T3 line every two to five square miles. You have to think about where to deploy, in three dimensions, and about the density of the client devices. Then there is geography and smaller physical obstructions like water, trees and buildings. And, you need to build for scale and for the eventual upgrade to adding voice and video. It’s very complex.”
The question of scalability in the face of an undereducated customer pool has risen to the surface for many vendors.
“People put up networks that cover a big area, but don’t know how much capacity to build for,” says Phil Belanger, vice president of marketing at BelAir Networks, which has expanded its product portfolio to include “every flavor of mesh,” he says, “from single-radio to high-end multiradio nodes. So we can design a low-cost network that is upgradeable to wherever they need to go. We educate customers on the benefits of multiradio, but can provide a single-radio option in the form of the BelAir50, where necessary to win the deal.”
While most mesh is Wi-Fi-based so widely deployed Wi-Fi clients can access the networks, most have proprietary protocols and extensions that make them unique in how they are deployed. There also are mesh networks built on proprietary, cellular and pre-WiMAX hardware. With spectrum allocation of major concern to operators and vendors alike, many are technology-agnostic.
Motorola Inc. rolled out MOTOMESH last month. This product is aimed at municipalities and other metro applications. The nodes include four radios that work with various networks and frequencies, including 2.4GHz for Wi-Fi and the new 4.9GHz band dedicated to public safety.
To overcome security and performance concerns for municipal applications, Motorola introduced proprietary mesh-enabled architecture (MEA) radios, which provide mobile connectivity at speeds of up to 250 MPH; separate frequencies for sensitive homeland security, public safety or municipal applications; and location-positioning that doesn’t rely on GPS and other features.
With a lack of consensus in an industry where standards are prized, the question remains: how far can mesh get.
“When it comes down to it, a mesh architecture makes any radio technology better,” says Rick Rotondo, director of marketing for Motorola’s Mesh Networking Product Group. “Due to physics, a mesh approach provides fewer bandwidth bottlenecks, it’s self-healing, provides multiple data paths and provides mobile benefits. This will be the norm in the future, regardless of the radio technology cities choose to use.”
more mesh
Most often, a mesh network is deployed using Wi-Fi technology, and while the underlying gear may be standardized, there are often unique software extensions and tinkerings that result in each company’s solution becoming customized and proprietary in its own way. The result? A splintered market where different companies’ solutions can’t talk to each other.
Seeing potential downsides to this reality, the IEEE 802.11s Task Group is pushing to standardize Wi-Fi-based mesh wireless LANs around the world. Benefits would include equipment interoperability and a seamless experience for wireless users, ease of deployment, the promotion of regulatory consensus, lower equipment costs, and the ability of a carrier to offer a managed service.
Out of 15 proposals for a new global mesh WLAN standard presented at a summer 802.11s meeting in San Francisco, two are considered frontrunners for adoption. A proposal from The Wi-Mesh Alliance, spearheaded by Nortel, builds on current 802.11 or Wi-Fi protocols to maximize backward compatibility and reuse of networks. A competing proposal from Intel Corp., Motorola Inc., Nokia, NTT DoCoMo USA Inc. and Texas Instruments Inc. is the other finalist. This one is known as SEEMesh — short for Simple, Efficient and Extensible Mesh.
“We want to address three main usage scenarios with the Wi-Mesh standard,” says Bilel Jamoussi, director of strategic standards at Nortel, “and those are the consumer electronic or SOHO user, the security and military arena and metro or campus deployments.”
Rather than requiring extensive overhaul to existing networks, the proposed architecture will be layered on top of existing mesh scenarios, performing a “mesh coordination function,” he says. It includes extensions to 802.11i to secure traffic between access points, not just from client to node and an extension to 802.11e for QoS to enable multimedia applications. Other aspects of the proposal include measurements of radio capabilities like power and channel availability, to allow for automatic resource management for enhanced spectrum spatial reuse, channel access coordination and RF resource management solutions.
A lightweight routing protocol will enable resilience, and a topology database will provide intelligent decision-making for unicast, multicast or broadcast traffic. The design supports both single-radio and multiradio platforms, and the proposed Medium Coordination Function offers three modes of operation, which allow for simple and robust implementations as well as for sophisticated solutions offering optimal performance and spectrum efficiency.
“Secure, scalable and self-configuring,” says Jamoussi. “The final specification needs to address these adequately in all three market segments.”
As for details on SEEMesh or its differences from the Wi-Mesh proposal, its backers aren’t talking.
“We are in the process of finding commonalities between the two and vetting the differences in approaches,” says Jamoussi. “The goal in the next few meetings is to take the best parts of both and converge on a single document.”
By mid-2006 the task group should have a final proposal, with the gelled specification likely in 2007 or early 2008.
| Links |
| BelAir Networks www.belairnetworks.com Intel Corp. www.intel.com Motorola Inc. www.motorola.com Nortel Networks www.nortelnetworks.com NTT DoCoMo USA Inc. www.docomo-usa.co Strix Systems Inc. www.strix.com Texas Instruments Inc. www.ti.com Tropos Networks Inc. www.tropos.com The Wi-Mesh Alliance www.wi-mesh.org |
