Wireless Integration - The Lowdown on LMDS

Wireless Integration Xtra! ImageMap - turn on images!!!

Carrier Services, July/August 1998

The Lowdown on LMDS

By Dan Sweeney

LMDS (local multipoint distribution services) is a new type of stationary broadband wireless access technology designed for a mass subscriber marketplace. Based on millimeter microwave frequencies - 24GHz and above - LMDS now offers a potential for cheaper in-building bandwidth than newly installed fiber or copper. In fact, it's the microwave technology that could make wireless data for the enterprise both cheap and ubiquitous.

Now that FCC auctions for the LMDS spectrum are complete, enterprise buyers will see LMDS "first mile" network services becoming available by mid-1999. Alternative carriers will tout LMDS as the ideal candidate for integrated voice and data in new multitenant buildings. With just an antenna at the end-user site, located within five to ten kilometers of an operator's two-way LMDS cell, the access technology can deliver huge chunks of bandwidth. Corporate applications can include anything from Ethernet and Fast Ethernet LAN extension to full-scale campus networking, alternative cable TV, video conferencing, high-speed Internet access and corporate video on demand.

Unlike privately maintained wireless bridges, LMDS entails no major equipment outlays on the part of the subscriber. Pricing for the service has yet to be established in any significant number of markets, but most analysts predict that LMDS monthly charges will amount to a fraction of T1 pricing for equivalent bandwidth - pricing comparable to competing Digital Subscriber Line (DSL) phone services and cable-TV networks.

"The cost of hardware for enabling data is roughly the same for LMDS and its chief competitors, cable and DSL," says Weston Vivian of Vivian and Associates, a consulting agency in Ann Arbor, MI. "This means that market pressures will determine pricing for the services," so that LMDS connections are apt to be priced at below $200 apiece - an alluring prospect for those accustomed to T1 rates.

In short, LMDS could represent a significant threat to incumbent local exchange carrier (ILEC) leased line services. For enterprise network buyers, this is great news: LMDS is a competitively priced wireless alternative for data that gives away nothing in speed to any terrestrial landline network with the exception of fiber-to-the-building.

LMDS basics

An LMDS public network is analogous to a campus wireless LAN or to the Metricom metropolitan wireless "MAN" (medium area network). The Metricom networks, currently operational in only a handful of cities - most notably San Francisco and Seattle - are somewhat analogous to LMDS in that they are wireless and use a cellular network architecture, though speed, at 164kbps max, is considerably less. Like these older networks, an LMDS system consists of a series of cells whose centers are defined by individual base stations, and of a central control point to which all of the base stations communicate.

Unlike these earlier types, the LMDS network uses highly directional antennas - sectorized antennas at the base station and single-beam parabolic microwave reflectors at the subscriber site. Subscriber units closely resemble the small dish satellite receivers used to pick up Direct Satellite broadcasts, while base station antennas tend to follow the same general form as those employed in cellular and PCS telephone systems.

LMDS transmissions are strictly line of sight. "Even foliage can interrupt transmission," notes Bill Frezza of Wireless Computing Associates of Yardley, Pennsylvania. For this reason, carriers are apt to target business districts where rooftop mounting of subscriber dishes is permissible.

Dave Ackerman, executive vice president for Winstar, a nationwide licensee in the 38GHz and 28GHz regions and a pioneer in developing multipoint microwave feels that critics make too much of line-of-sight requirements. "You get around line of sight by the placement of the hubs," he explains. Winstar has aggressive plans to deploy LMDS in both businesses and residences, especially new construction and multitenant buildings.

Big spectrum

LMDS services are permitted at a number of frequencies - 24GHz, 28GHz, 31GHz, 38GHz, and, soon, 40GHz. The 28GHz region has the most generous spectrum allotment, 1.3GHz, and currently offers the greatest potential for bundling of diverse services.

The capacity of 28GHz LMDS, which actually consists of three bands - 27.50-28.35, 29.10-29.25, 31.00-31.30 - is such that systems can accommodate high-speed Internet access. Other applications include multiple virtual private networks for corporations and government agencies, packet or ATM telephony and streaming video, including video broadcasting.

Originally, LMDS was conceived as a competitor to cable television, offering much the same kind of content to a consumer subscriber base. But Internet growth and demand for LAN extension services, whereby private LANs are linked via public carrier or wireless bridge with remote sites, have caused most bidders to revise their business plans. Today LMDS is seen primarily as a data pipe, though it needn't just be that.

LMDS licensees have announced plans for supporting all kinds of corporate network services, including secure file transfer and messaging within a virtual LAN context, video conferencing, and IP telephony. Contrast this approach with cable-television operators who stress merely one data application, namely, Internet access. (DSL suppliers, on the other hand, fall between the two camps: they've targeted consumer markets for high-speed Internet access and also corporate data applications such as LAN extension.) According to Dan Ernst of the Strategis Group (Washington, DC), data and telephony are apt to receive the most LMDS operator emphasis.

The marriage of ATM and LMDS

As operators roll out LMDS, they'll have the advantage of technology that can aggregate multiprotocol traffic into a convenient "wireless fiber" (as Winstar calls it). Their trump card is ATM; that is, most LMDS equipment utilizes embedded ATM protocols. These ensure that separate "virtual circuits" are assigned to each user, and that, within these virtual-circuit connections, LMDS allocates bandwidth according to circuit use at any particular time. Delays are managed to avoid perceptible degradation of service. For example, if the subscriber chooses to set up a video conference, adequate bandwidth will be guaranteed to ensure a predefined level of video resolution throughout the conference.

ATM is not used in most cable data systems, and while much work has been done in marrying ATM to DSL, problems remain in combining the two technologies due to the marginal throughput speeds of DSL over less-than-ideal copper connections. LMDS consequently stands alone among widely available, competitively priced high-speed data services. It's the only broadband access solution that uses protocols accommodating different types of users over shared bandwidth.

And now, the dark side

Attractive though it appears, LMDS is not without manifest drawbacks. Unlike its chief competitors, cable and DSL, it lacks any industry standards for interoperability, and thus won't currently support roaming or a consumer mass market for radio modems. In addition, much existing equipment is limited in its interface capabilities, not yet supporting Token Ring architecture, for instance. Finally, multipoint microwave service based upon a cellular network configuration is not fully mature.

In sum, though LMDS represents an attractive combination of speed, managed bandwidth, adaptability to a wide range of applications and, presumably, affordable pricing, it remains in an early state of development. In the next section we will consider what that means to prospective purchasers of high-speed data services.

Fat pipes at fat prices - the traditional approach

For many years the high-speed connection of choice for corporate enterprises has been the dedicated T1 telephone line or fractions or multiples thereof, leased from the local-loop provider. T1 connections are secure, reliable and consistent, providing 1.5Mbps throughput at low latency. T1s have not gotten any faster over the years, however, nor will they, and pricing is generally steep, averaging about $1000 per month and sometimes considerably more. If Ethernet speeds are required, managers resort to multiple T1 connections at some multiple of basic T1 pricing. Maximum speed available over leased-line circuit connections has generally not exceeded 45Mbps, less than half the throughput of a fast Ethernet LAN (100 Mbps.)

Fat pipes at slender prices - the emerging approach

Since the late 1980s, an alternative to the T1 connection has existed in ISDN (integrated services digital network), a higher-speed desktop phone connection over ordinary twisted-pair phone lines. ISDN, however, is considerably slower than T1. Basic ISDN over a single wire is 64kbps, while two-wire connections merely double that rate. Multiple ISDN connections provide 64kbps increments, and various forms of data compression can increase throughput by a factor of four. Historically, ISDN connections have averaged about $100 per month for two wires, but recently competitive providers have begun offering service for under $50 per month in some markets. A multiple ISDN connection may be more cost effective than T1 if throughput requirements are not exorbitant.

Recently many telcos as well as capacity resellers and CLECs have begun to offer new types of higher-speed data services competing with traditional leased-line solutions. These include frame relay, SMDS medium area networks and ATM services. Most of these utilize T1 or fractional T1 connections to a network access point, beyond which a signal will share bandwidth with other network subscribers. Such services are not necessarily cheaper than T1 when only two sites are linked, but where remote nodes are involved, they cost considerably less.

While these services normally operate over ordinary telephone infrastructure, employing both fiber and copper where appropriate, availability and cost vary widely. The requirement of multiple T1s for access ensures that overall cost will remain substantial, placing them at a disadvantage in this regard vis a vis LMDS. Note, however, that both ATM and frame relay can be adapted to an LMDS physical layer, so the distinction is really between fast relay over copper and fiber, and fast relay over the air.

Broadband cable

Within the last two years, cable television companies have begun to offer data connections with maximum throughput rates of several megabits per second, that is, multiple T1, at flat rates of well under $100 per month. While still only available in scattered geographical markets, cable data is now being promoted by all of the major MSOs (multiple system operators) including Time Warner/Media One, Comcast, Cox, Adelphia and others, and phased rollouts continue on a regular basis [see "Washington Wireless," a new column, page 11]. Within the larger metropolitan areas, cable data is likely to be ubiquitous by the year 2000.

Cable modems certainly offer bandwidth for the buck, but from an enterprise management perspective they tend to be problematic. First of all, specified throughput speeds, while undeniably impressive, are normally stated only for the lightest network loads. Most cable data systems have no provisions for guaranteeing bandwidth or controlling latency with increasing traffic, nor can they easily provide multiple private intranets. Moreover, cable operators themselves are mostly new to the data networking business, and many have little or no in-house expertise in managing networks. Finally, many lack facilities in business districts and are loath to construct them.

Most cable data services heretofore have consisted of fast Internet access for consumers plus assorted online programming packages along the lines of America Online's offerings. Those few cable operators who do pursue business markets tend to specialize in telecommuting applications rather than providing true WAN connectivity. In other words, the services tend to fit the mold of broadband residential rather than the corporate LAN/MAN/LAN connectivity model. Further, significant numbers of cable data providers only offer one service, that is, the return path consists of a telephone connection only - a major limitation because IP transmissions in one direction require periodic acknowledgments to be sent back to the sender. A slow return path will limit the speed of the forward path as well - a crippling limitation in an extended LAN.

In short: well-engineered, meticulously maintained cable systems can pass data successfully, but their basic infrastructure is simply not designed for that purpose, nor are most adaptable to two-way data traffic without extensive modifications and improvements. In terms of sheer performance, LMDS systems may work better because they're specifically designed for data and broadband transmission.

DSL

DSL (digital subscriber line) is the name for a family of technologies for extending the bandwidth and throughput capabilities of ordinary twisted-pair copper telephone lines. In simplest terms, a combination of signal amplification and digital phase and amplitude equalization is used to permit twisted pairs to transmit at frequencies many times higher than those normally supported.

While the signal processing used in DSL is adaptive, the technology can only go so far in overcoming copper losses. Highly corroded cabling or lines encumbered with taps or loading coils will defeat even the most aggressive applications of the technology. Unfortunately, for the proponents of DSL, much of the telephone infrastructure in the U.S. is old and deteriorating, and thus unsuitable for DSL without extensive reconditioning.

DSL throughputs vary, not only from the condition of the plant, but according to the type and specifications of equipment used. HDSL, the speed demon of the family, can push 50Mbps over a few hundred yards of virgin copper, but the significantly slower ADSL is apt to be much more widely deployed. Most local loops are not expected to support more than a few megabits per second - on a rough parity with cable.

All of the RBOCs are in the process of rolling out DSL services, and while subscribers lag behind those for cable data, DSL is definitely the comer among affordable high-speed wireline data services. In many ways, DSL is better suited to the business user than is cable. As with T1, the circuits are nailed up and normally not shared up to the central office, which means that throughput rates are relatively invariable with network loading. Privacy and security are greatly enhanced, and there's no video to hog bandwidth.

Worth waiting for?

In terms of network availability, T1 and ISDN are ready today, while cable and DSL are imminent or widely available the day after tomorrow. LMDS is the near future. By year's end, cable data will be popular and DSL available at selected network sites. LMDS will be offered in beta in at most a few metropolitan markets and possibly none for full-scale deployments this year. Currently none of the licensees offer anything but beta services.

Still, spectrum has been auctioned and numerous companies are licensed to begin service. LMDS cannot be said to be speculative in the sense of Third Generation cellular or LEO two-way satellite service. LMDS is imminent.

Will it provide solutions for cash-strapped MIS officers? Perhaps the best assessment is given by an admitted competitor whose company ultimately decided to embrace the technology.

"We're firmly committed to cable data," says Steve Craddock, vice president of new media development at Comcast, "But we studied wireless broadband and decided to buy some spectrum. It looks good to us."

Back to the Carrier Services Index

Back to MOBILEINFO.COM's site.

The Communications Web Sites

Visit the Optoelectronics Web Sites
Meet the Publisher

Comments or Questions about this site, please contact the Webmaster.
�Copyright 1995-1998 PennWell Media Online

This page was updated on 11-Jan-99.