Wireless genie out of bottle: open source WiFi here to stay.

For the past several years, fixed wireless has been through numerous trials--and has failed to catch on. Originally conceived as a relatively easy way to bring broadband Internet access to rural areas under-served by cable and DSL service, most RBOCs, as well as the nation's long distance carriers, have also tested the service in metro areas (Tucson, Atlanta, and Fairfax, Va.) and found technical difficulties, scalability issues, and high costs.

Now, a new movement is afoot that may change the economics of--and the technology behind--fixed wireless.

Unfortunately for providers, it relies on open source software and is intended to be free. At the same time, however, new efforts aimed at standardizing wireless Internet access in the licensed spectrum are gaining momentum. The wireless genie, it appears, it out of the bottle. Like it or not, there's no going back.

To understand the origins, features, and possible limitations of broadband wireless technology, it is helpful to know how traditional fixed wireless works. In its original incarnation, fixed wireless service is a point-to-multipoint, aggregated bandwidth solution. A cellular tower transmits a very high bandwidth stream (generally 5 to 20Mbps but going as high as 400Mbps), which is then sliced up among end users, each of whom has a small receiver--generally a small dish or panel mounted on the roof. Different companies use different bandwidths, but most installations have utilized 2.4GHz or 5.8GHz. Local Multipoint Distribution Service (LMDS) technology has also been employed by many providers, and offers a range of about five miles.

Limitations

One of the main reasons why LMDS and other early fixed wireless solutions did not take off is that they required line-of-sight (or occasionally near-line-of-sight) connections: that is, the tower (or base station) needed to have an unobstructed view of the receiver. This solution proved of limited use in hilly/mountainous areas of the country, as well as in cities, where multiple towers were needed because of skyscrapers.

Last year a few providers, including Verizon and Sprint. began trials of fixed wireless technology that does not require line-of-sight. Verizon's equipment, supplied by BeamReach, uses Adaptive Multibeam orthogonal frequency division multiplexing (OFDM) technology, with data rates of about 1.5Mbps (per customer) and a range of about five miles. The technology is not intended to replace DSL. But because DSL only has a maximum usable distance of about 18,000 feet (or 3.4 miles) from the telecom Central Office, it is not available to all customers (and even some within the acceptable range are kept out of DSL's reach because of line quality issues).

Sprint partnered with Navini Networks and IP-Wireless for trials of each company's technology, the former's based on 2.4GHz and 2.6GHz Multichannel Multipoint Distribution System (MMDS) frequencies, the latter on Universal Mobile Telecommunications System and Code Division Multiple Access cellular technology.

Few Companies, Fewer Customers

While current fixed wireless solutions may be appealing to small niches (large corporate customers without T1 or T3 lines, those out of the reach of DSL and cable), they remain costly and with limited infrastructure to support them. The list of companies that have tried and failed with--or tried and ultimately abandoned--fixed wireless reads like who's who (and who's that?) list, and includes AT&T, WorldCom, Metricom (creator of the popular but unsuccessful Ricochet network), Teligent, WinStar, Hybrid Communications, and many, many more.

Part of the problem is that the expense of building and maintaining fixed wireless networks demands that the networks themselves be multi-modal. In the current economic climate, carriers are not in a position to create overlapping wireless networks that use different topologies and protocols. And, to make matters worse, the carriers' customers are now handling convergence in-house, combining divergent networks into a single entity that provides all necessary communications functions, including wireless and wired voice, and wireless and wired data.

"For the last several years the promise of collapsing multiple corporate voice and data networks was promised, but rarely realized," Daryl Schoolar, a senior analyst for market watcher InStat/MDR's Advanced Carrier Services, said in a recent research note. "While some service providers claimed to deliver a full end-to-end corporate solution over one network, it was often more marketing hype than technological reality. Now, however, corporations are starting to take matters into their own hands," Schoolar added. "Corporations are taking advantage of improvements in VoIP and IP-VPN to develop their own in-house unified voice/data networks, relegating service providers to being only a dumb pipe provider."

Such developments mean that carriers need to create value by building networks that offer a full range of services, not by simply offering a menu of a la carte items. "Despite constraints on capital budgets, service providers need to pursue these markets," Schoolar noted. Already, several well-known wireless OEMs--including Motorola and Proxim--are readying equipment that allows enterprises to install and manage their own VoIP/WiFi/cellular gateways (see last month's issue, available at wwpi.com, for a full discussion).

Access for All

It was perhaps inevitable, then, that the open source software movement would begin to train its sights on wireless IP networks. An umbrella organization called Community Wireless.org is attempting to popularize the somewhat novel concept that broadband wireless does not necessarily have to cost money. Here's how the concept works.

First, I pay for a broadband connection, say a cable modem with 2Mbps of bandwidth. I then connect my cable modem to a wireless router with an 802.11 access point (AP). So far, all I've done is to begin building a wireless network in my house. But, as most people know by now, it requires some effort, and a bit of technical savvy, to prevent wireless transmissions from traveling beyond the walls of a house. In fact, it's relatively easy to piggyback on a neighbor's unprotected wireless network, should one choose to do so.

The industry has stressed that it is crucial to protect your wireless network from hackers and outside intruders. But what if you didn't care about others using your network? What if, in fact, you freely allowed anyone within range, and with the proper hardware, to access the Internet through your connection? After all, very few of us use anywhere near the bandwidth available on our connections at any given time. Why not share the wealth?

In fact, this is what small communities of users have been doing, somewhat stealthily, all over the country for the past several months. "Most of the activity in the wireless broadband space has been in the license-exempt spectrum," In-Stat's Daryl Schoolar said in an interview. "Companies have been taking the 802.11 kernel and building on it for outside use. We estimate that there are already 1,500 to 1,800 small ISPs offering last mile wireless."

Now, an inventor named Jon Anderson in the U.K has created a device called the LocustWorld Hardware MeshAP that he specifically designed to access the Internet and distribute bandwidth, for free. According to Anderson's website, the MeshAP is a self-contained, Linux-powered device designed to provide mesh connectivity, either by itself as a mesh repeater, or as a client/server internet gateway to the mesh. You simply plug the device into a PC and it connects you into the wireless network. With an Internet connection, you can share your access via the device. Without one, you can access someone else's for free using the MeshAP. It also contains all the standard functions of a wireless access point, so it can be used standalone or as part of the network. Each unit has software that automatically sends new updates over the air. Anderson sells the device for cost (about $340).

"LocustWorld is a very unique company," says Gemma Paulo, a senior analyst with the Data and Voice Networking Group at In-Stat/MDR. "The MeshBox device extends the MeshAP into a set-top box Internet appliance and wireless mesh router. The MeshAP has wireless mesh routing software built into it, allowing for MeshAPs to form community networks, if you will, theoretically allowing for a much broader coverage for a fixed wireless service provider.

The success of such a system, of course, depends both on ubiquity and on license-free software. The latter is easily handled by Linux. The former is still in its early stages, since 802.11 is still relatively new. However, there are already Internet registries where users willing to share their connections can list their locations. And, with the proper hardware and the right conditions, WiFi networks can have a range of tens of miles.

There are other barriers, too. Users must be willing to offer their paid-for bandwidth for free. The bargain is simply a barter: I'll give you access to my network, as long as you give me access to yours. While many technophiles have been willing to make such a deal, the average user might not. And, of course, the user needs to be technically adept enough to protect personal data and private files from unscrupulous piggybackers.

There is one more problem, and it is perhaps the most significant. While individuals may benefit from free broadband wireless access, the companies who supply said access can quickly find their networks bogged down with unexpected traffic. Most cable and DSL providers explicitly prohibit hosted servers, and a shared wireless AP is in many ways even more bandwidth intensive. There is a good counter argument, of course. If I am paying $50 a month for access, that means I am entitled to as much access as I can use. If I can't use it all, I am free to share it with others. This, after all, is the concept behind home networking. Community wireless simply brings the concept to its logical conclusion.

The big cable providers are having none of it. Their usage policies state that use of the signal beyond the boundaries of the home is piracy, and may result in service termination. Of course, since the unauthorized access is provided through a router at the customer premises, determining whether a broadband connection is being shared can be difficult. (Though not impossible: Allot Communications sells software that helps providers track how their broadband connections are used--or abused.)

The larger problem is that supply cannot keep pace with demand. Consumers want wireless net access, but the big RBOCs are stuck in trials of broadband wireless, mostly because they have to deal with scalability, terrain, and cost issues that smaller, local wireless ISPs can ignore or easily work around. "RBOCs are interested in this technology, especially to extend DSL service," says In-Stat's Schoolar. "But Sprint and WorldCom hold the majority of equipment that works in the licensed spectrum." Small ISPs know where each tree and hill in their territory is, Schoolar adds, and can tweak their networks as needed. The bigger providers don't have such flexibility. "It's hard to imagine an engineer or marketing manager from Sprint going into a neighborhood to map all the hills," he notes.

Broadband Aid

The monkey in the wrench in this broadband access free-for-all is an emerging standard for broadband wireless known as 802.16a, which was approved by the IEEE in February. 802.1 6a is an amendment to 802.16, a wireless Metropolitan Area Network (WMAN) technology that has been winding its way through the standards process for a few years.

Like other fixed wireless technologies, an 802.16-based signal is beamed from towers and uses OFDM at its core. There are no line-of-sight requirements. 802.16 was originally intended as a backbone-class, high frequency (1066GHz) standard for last-mile enterprise transport in areas where laying cable was too expensive or impractical, or to connect multiple optical networks. But with the explosion in popularity of WiFi as a local wireless topology, many companies have seen an opportunity for a broadband, wide-area wireless standard operating at lower frequencies.

The 802.16a amendment will support a frequency range of 2 to 11 GHz, making it more suitable for consumer premises equipment. Not surprisingly, 802.16 has its own industry consortium, called the WiMax Forum, that performs interoperability tests of the efforts that come out of IEEE working groups. While WiMax membership is still small, a few big names are in evidence, including Agilent, Nokia, and Hughes Network Systems. (The organization brings to mind the early standardization efforts of WECA, the Wireless Ethernet Comp-atibility Alliance, which morphed into WiFi.) 802.16 may, eventually, be the standard that allows consumers to roam between WiFi access points while maintaining their connections.

Will 802.16 be a boon to licensed or unlicensed wireless access technologies? Some observers say both. While 802.11 connections are. being used beyond their intended distances with some success, such solutions can be proprietary, expensive, and insecure. "802.16 will affect both licensed and license-exempt providers by increasing interoperability and bringing cheaper silicon," among other ways, says Daryl Schoolar.

While we are unlikely to see 802.16 products before the end of the year, there's little doubt in the industry that they will arrive. But much depends on what the FCC does with unused MMDS spectrum, Schoolar says. "If the FCC opens the spectrum to other MMDS providers, and vendors can supply adequate MMDS gear, MMDS could push license-exempt back to the fringes of the market," Schoolar posits. "If the FCC opens up some MMDS spectrum for license-exempt wireless ISPs, that should spur further market development."

Regardless of the timing of products and the FCC's decision, the incredible success of WiFi has clearly demonstrated to the industry that consumers are ready for a simple, low-cost way to access the Internet without wires. If the big boys won't provide an economical way to do so, someone else will surely blaze the trail for them.

www.agilent.com

www.allot.com

www.communitywireless.com

www.ipwireless.com

www.locustworld.com

www.motorola.com

www.navini.com

www.proxim.com

www.sprint.cm

www.verizon.com

www.wimaxforum.com