Wireless ATM flexes its muscle

Progress toward melding ATM with wireless technology is currently limited to use in the backbone and for connecting mobile switching centers. However, future applications will deliver ATM directly to the end user

The definition of wireless ATM varies considerably depending on who's talking. But the one thing on which pundits agree is that the fusion of wireless and asynchronous transfer mode technologies exemplifies a fresh archetype for offering streams of voice and data information, and it is leading the way to enhanced high-speed data and multimedia communications.

To some, wireless ATM is the use of wired ATM platforms as backbones for digital wireless switching networks. Others say true wireless ATM requires actual integration of ATM with wireless network management for bandwidth allocation. Still others feel the concept entails an end-to-end wireless ATM solution that includes wireless ATM devices for end users.

ATM organizes information into 53-byte cells, each of which constitutes a fixed-size frame consisting of a 5-octet header and a 48-octet payload. ATM uniformly handles all transmissions-voice, data and multimedia-allowing a single network platform to deliver multiple services.

Despite struggling for the past decade to find its niche, ATM is experiencing success as the backbone for various networks, whether wide area or campus area, wireless or wireline. In fact, 60% of Internet protocol traffic crossing the Internet is going over ATM backbones, notes George Dobrowski, president of the ATM Forum. And the demand for high-speed Internet access and neoteric data-rich communications is driving ATM's growth in both the wired and wireless worlds.

Still, ATM deployment has a long way to go before the technology proves its strength. "People recognize that ATM is a good core technology, but it's not possible to replace everything overnight," says Dipankar Raychaudhuri, assistant general manager at NEC USA's C&C Research Lab in Princeton, N.J. The cost of bringing ATM to the desktop remains high, although it is falling rapidly, he says.

"ATM is extremely cost-effective in terms of being able to handle large bandwidth multimedia streams," he says. "If you look at a small ATM switch today, such as NEC's Model 5S, it can handle 2.4 gigabits of data for just $20,000."

Building the hybrids One new growth area for ATM is in creating a backbone for nascent digital cellular and personal communication services (PCS) networks. Lucent Technologies is marketing ATM backbones for networks based on code division multiple access (CDMA) technology. The company's GlobeView 2000 ATM switch frees carriers from having to establish circuit-switched or dedicated, fixed connections between mobile switching centers (MSCs).

A key aspect of CDMA is its soft handoff capability, which simultaneously connects callers with up to three base stations while a frame selector decides which transmission frames to deliver to the network. Soft handoffs generally function between minicells served by one MSC, but Lucent claims its ATM technology extends the soft handoff capability to function between adjacent MSCs (Figure 1). "In order to do the soft handoff with minimal facilities cost, you put the ATM switches in so you can connect all the switches together," says George Foley, market development director for Lucent's data networking business unit.

Carriers that implement such an ATM backbone get the side benefit of being able to fold in operations support systems, Foley notes. He says a half dozen CDMA networks worldwide have implemented Lucent's ATM solution.

Cisco Systems also has marketed its ATM technologies for cellular and PCS backbones, but Junaid Islam, Cisco's group manager for the wide area network business unit, disputes the perception that such a scheme constitutes true wireless ATM. "That is ATM in the wireless world," he says. "Wireless ATM is where ATM is used as the primary bandwidth management scheme for spectrum." Such bandwidth management is essential to Cisco's ATM solutions for metropolitan area networks (MANs).

"A lot of alternative carriers don't want to use a [Bell company's] fiber because the cost base is too high. So they are very interested in using microwave or satellite technologies," Islam says.

However, such access modes often rely on simple time division multiplexing (TDM) in which bandwidth is dedicated for different channels or services. But alternative carriers now can integrate ATM to statistically multiplex traffic at broadband speeds, he says. "That is the big story regarding where ATM's biggest impact is right now," according to Islam.

Integrated microwave- or satellite-ATM networks are becoming especially popular in less developed countries where leasing T-1 or E-1 lines is cost-prohibitive, he adds.

Cisco has invested in Netro Corp., a Santa Clara, Calif., company that also targets MANs, local loop and last-mile customer access networks with its AirMAN wireless ATM product family. The company's point-to-multipoint and point-to-point broadband solutions integrate network management software and value-added ATM and frame relay services.

"People look at wireless ATM as a service unique to wireless carriers, and it's not. It's just one method of transmitting data that is more efficient than TDM," Islam adds.

Reaching the users Wireless ATM developers are looking beyond current breakthroughs to a day when ATM can be delivered directly to the end user. Using his definition of wireless ATM, Cisco's Islam notes that in the short term, "ATM over microwave or satellite systems are the current plays, while ATM to the mobile user is the long-term play."

Developers are working to deliver ATM to the end user in the 5 GHz spectrum, some of which is allocated in the United States as the National Information Infrastructure (NII) band and in Europe as the HiperLAN band. The ATM Forum's wireless ATM working group is adapting ATM standards for use in the NII's 5.2 GHz spectrum.

Although the NII band is reserved for unlicensed, low-power applications, wireless carriers should consider capitalizing on the NII's expected popularity by offering their subscribers dual-mode, dual-frequency products, says Dobrowski, who in addition to his role as forum president is director of broadband switching and signaling technology at Bellcore.

The European Telecommunications Standards Institute is working on specifications for the NII band's radio physical layer, while the ATM Forum is adapting the ATM protocol for use in the band. The forum is addressing bandwidth management, signaling protocol and mobility issues that would let users carry the same device at the office, home and while traveling. Since many countries are agreeing to free up the same or similar spectrum, such a device even could be used in international locations. This ubiquity would allow users to get the same services and applications wherever they are, thus encouraging competitive bundling and pricing of services, Dobrowski says.

He expects to see trials of 5.2 GHz equipment in 1998 and commercial devices in 1999. Devices might run the gamut from PC cards delivering high-speed data for personal computers to small hand-held devices outfitted with LCD screens for videoconferencing capability.

NEC's C&C Lab has been developing the wireless ATM concept for several years. "In end-to-end wireless ATM, you transmit ATM cells on both the radio and the switching network. Mobile ATM is the subset [in which] you transmit ATM cells in the switching network so the radio access technology could be whatever you want," Raychaudhuri observes.

The company created WATMnet to deliver multimedia information in the NII band in the United States and in Europe's related HiperLAN band. However, NEC also has moved wireless ATM into the nomadic environment with its mobileATM switching network, which includes ATM base stations.

The ATM base station concept could be used to facilitate global wireless roaming by enabling carriers to serve multiple air interfaces. With NEC's ATM base station, the radio protocol is terminated at the base station and the generic mobile ATM service is used to move mobile data around, enabling the network to accommodate any radio protocols the carrier chooses, Raychaudhuri says.

Meanwhile, numerous joint European projects are focused on hybrid wireless ATM networks. One in particular, called Magic WAND (for Wireless ATM Network Demonstrator), is targeting 5 GHz-band multimedia communications direct to end user devices. The group is focused on developing technology for use in fixed locations or campuswide networks as opposed to high-mobility, vehicular networks such as cellular and PCS.

"Campuswide networks require mobility, so there's a possibility to extend [wireless ATM] service into a wide area if you have a really densely populated area. But I do not think ATM will still stretch into today's cellular or PCS technology," says Jouni Mikkonen, R&D technical manager at Nokia Mobile Phones in Finland and Magic WAND technical manager. High mobility presents technical problems for end-to-end wireless ATM, Mikkonen says, and the finite spectrum afforded cellular and PCS limits any broadband applications.

Yet another European project-called the System for Advanced Mobile Broadband Applications (SAMBA)-is aimed at providing transparent ATM connections and transmission of ATM cells up to 34 Mb/s in a broadband cellular radio environment at 40 GHz (Figure 2). Trials demonstrating services such as a wireless TV camera and medical applications are planned for Expo '98 in Portugal. The SAMBA group is promising successful handoffs and location management for the reliable transmission of ATM cells. Part of SAMBA's goal is to highlight possible applications in next generation broadband wireless networks.

Eyeing LMDS Another prime market for wireless ATM is local multipoint distribution service, which is expected to provide a variety of two-way voice, data and video services. The Federal Communications Commission is slated to auction 984 licenses in the 28 GHz and 31 GHz bands beginning Feb. 18.

Germany's Siemens AG is one of several companies targeting ATM solutions for LMDS. The company's Public Communication Networks Group has joined with Newbridge Networks of Canada to create the MainStreetXpress 36190 ATM switch, which can be integrated with wireless equipment from BNI of Winnipeg, Manitoba, to produce a wireless MAN (Figure 3).

"We are enhancing actual broadband wireless with ATM connectivity to an ATM backbone network, which is strategic in offering a multiservice platform in the wireless environment," says Philip Yim, Siemens' U.S.-based manager of broadband product line management.

Likewise, Cisco has talked with partners about entering LMDS with a wireless ATM solution. "The idea is to have Cisco [customer premises equipment], somebody else's wireless package and maybe a Cisco backbone," Islam says.

The future of wireless ATM, regardless of how it is defined, is full of possibilities but few absolutes.

Many feel that ATM will be integral to next generation broadband wireless as envisioned in the International Mobile Telecommunications-2000 project being undertaken by the International Tele-communication Union (Figure 4).

"One, ATM allows you to carry more voice [signals], and two, it is capable of bursty data. It also allows wireless carriers to use public services for backhaul, which is totally new. Right now, if you're a wireless carrier, you have to build a huge T-1 mesh. If you can use someone else's network, it's a lot cheaper," Islam says.

Raychaudhuri also sees a role for ATM in the next generation. "In the future, as carriers begin to offer broadband services, they can just plug a 5 MHz broadband radio card into the [ATM] base station and then start offering broadband services with the same infrastructure."

Likewise, the ultimate goal is to bring wireless ATM to the end user. "You could use a wireless ATM gadget to hook up to the office LAN at high speeds, or use it to make a wireless LAN into a wireless MAN once a user steps outside the office," Islam notes. "That is a more futuristic approach: I see that as Phase 5. ATM over microwaves or wireless may be Phase 1 or 2. We will have to go through many iterations before we get to that true wireless ATM to the user. Using ATM as the backbone is a today story. Getting wireless ATM to the end user is a five- to 10-year story at best," he adds.

But others see ATM to the end user occurring much sooner. "What I predict in the future is that these [wireless ATM] devices will be dual-mode," says Dobrowski. "In my discussions with all of the classical wireless suppliers, they are working on next generation designs that will be dual-mode. You'll be able to use the 5.2 GHz spectrum while you're in the office, at home, in a hotel or in a conference room, and then you'll be able to switch to the other mode-the more classical cellular mode-while you're in the car driving to the office."

Some companies such as Boeing are working on specifications that would enable the use of 5.2 GHz devices on aircraft via satellite transmissions, he adds. "The Boeing folks have told me that their next generation aircraft will be based on ATM. In fact, they'll have ATM right to the seats along with power for your laptop."