Deploying Point-to-Point Wireless Links

Markets and Standards

The broadband fixed wireless market can be segmented in several ways, most notably between the point-to-point and multipoint arenas. The dream of the multipoint industry is to provide an alternative to DSL, scalable to deliver service to thousands of subscribers in an area. Unfortunately, vendors have struggled to issue a compelling story, primarily because implementing multipoint systems is technically challenging, particularly when line of sight to each subscriber is required.

The point-to-point market shares many technical challenges with multipoint, but the economics are much different--the aim is to save thousands of dollars per month on service charges as opposed to competing with $79 DSL. And unlike multipoint systems, point-to-point systems make radio interference issues manageable because you don't have hundreds or thousands of links communicating to the same base station.

In multipoint systems, standards are critical because they drive cost points lower and allow multivendor equipment interoperability. But in the point-to-point segment, standards aren't quite so important because you'll typically deploy turnkey systems. And though vendors' offerings may be proprietary, the industry has enough competition to encourage innovation and aggressive pricing.

In today's market, purchasing a point-to-point system based on 802.11b may provide the lowest up-front cost, but you'll take a performance hit for the overhead that comes with adapting a LAN standard to the point-to-point market. Further, the technical features of these products may complicate your efforts to deliver rock-solid reliability.

Technical Points

The first point-to-point wireless systems we tested in our Real-World Labs® in 1995 (see "The Bridges of Wireless County,") gave us a feel for the challenges associated with field deployment of links requiring line of sight over distances of 3 miles. Much has changed in the past seven years, and redesigned systems are now a lot easier to install, though the requirement for line of sight between antennas remains. While you may enjoy initial success in deploying point-to-point wireless even if you don't understand the technical issues, you'll sacrifice long-term reliability. For example, we've seen point-to-point systems deployed using inappropriate antennas, and while they work on Day 1, their susceptibility to interference may cause them to fail later, probably when the installer is cruising the Caribbean.

So what are the core technical issues involved in engineering links that have very high reliability? First, you need to understand some of the basic physics associated with analog radio systems. Second, you need to be aware of the governmental regulations designed to facilitate the shared use of radio spectrum, particularly in the unlicensed 2.4-GHz and 5-GHz bands. And finally, you need to recognize the design trade-offs that vendors make when developing systems and the consequences for specific instal- lation scenarios.

That Old Math Magic

For several years, we have referred to the "magic of physics" associated with wireless systems. Obviously, there's nothing magical about it, but the underlying complexity of radio engineering makes it difficult for nonexperts to grasp how hundreds of megabits of data can be pushed down an invisible pipe every second. Regardless, you'll need to understand some basic principles to install one of these systems.

First, there's frequency. Radio-wave behavior varies with changes in frequency. As you move to higher frequencies, the wavelengths become shorter, affecting their propagation characteristics. For example, low frequency AM radio waves (500 KHz to 1,500 KHz) can propagate for hundreds or thousands of miles, and adverse weather and physical obstructions have little effect. But at frequencies greater than 40 GHz, radio waves are readily absorbed by rain and fog. Most point-to-point systems are deployed at microwave frequencies, with operations in the unlicensed 2.4- and 5-GHz bands most common.

Closely related to frequency is the amount of bandwidth available for a specific link. This is a product of both the size of the RF channel (for example, 10 MHz) and the analog-to-digital conversion techniques employed. One key issue facing system designers is spectral efficiency, or the amount of data that can be packed into an RF channel. Since available spectrum is limited, it's obviously beneficial to use advanced signal modulation techniques to pack as many bits into a channel as possible. However, as spectral efficiency increases, so too does the cost of underlying components and the possibility of errors. It's a delicate balancing act. Most point-to-point systems can operate at multiple alternate channels within a specific frequency band, which can be important in mitigating interference.

While it's common to refer to the power of RF systems in watts, designers of point-to-point systems more often use decibels (dB), a logarithmic transformation that lets us measure the ratio between input and output signals while avoiding numbers with many zeros. One of the primary purposes of a radio system is to amplify a signal, thereby generating gain, measured in decibels. Without getting into the underlying math, you can understand a lot just by knowing that a 3 dB increase in gain means the signal is 2 times bigger. Radio power is often measured in watts or milliwatts, so when decibels are used to represent the ratio between input and output signals, using milliwatts as the reference (indicated by dBm), 0 dBm is equivalent to 1 milliwatt, 10 dBm to 10 milliwatts, 20 dBm to 100 milliwatts, and 30 dBm to 1 watt

Gain in point-to-point systems is a product both of radio output power and of an antenna's ability to focus that power. All other things being equal, higher gain systems are more difficult to build. Because all RF signals experience predictable loss over distance, you might logically assume those that operate over longer distances will need more gain. However, transmission is only one side of the coin. A system's receive sensitivity is equally critical. And as you probably guessed, more sensitive receivers are most costly to build.

In deploying a point-to-point system, you'll often be working with a loss budget. Loss in RF systems begins at the output of the radio transmitter and continues through to input of the receiver. Free space loss, also known as path loss, occurs as signals pass through the air, mainly because radio signals spread over distance, somewhat like water exiting a garden hose. As frequency increases, so too does path loss, meaning a 2.4-GHz system will have a greater range than that of a 5-GHz system of equal power output. For a frame of reference, note that a 2.4-GHz radio signal will experience a free space path loss of about 120 dB over a distance of 5 miles. (For a free space loss calculator, see "Free Space Loss.")

Loss also occurs in other areas, most notably in the cabling that connects the radio transmitter to the antenna. By integrating the antenna and radio into a single weatherproof outdoor unit, some vendors eliminate this loss.

As noted earlier, antennas provide gain in a radio system, but the way they do this is different from the way radios deliver gain. Radio-system engineers often refer to an isotropic antenna, a theoretical concept that shows the spherical nature of how signals would radiate from a single point in space. You can't build an isotropic antenna, but some designs come close. Omnidirectional dipole antennas (rabbit ears), for example, radiate in a pattern that looks like a doughnut with a very small hole at the center. Omnidirectional antennas generally provide modest gain and are inappropriate for point-to-point links.

Directional antennas provide additional gain by focusing the radio energy in a directional beam. They don't add any power to the radio signal, but they concentrate it, thereby increasing range relative to the theoretical isotropic antenna. The gain of an antenna relative to isotropic is expressed in dBi. There are many different directional antenna designs, but the most common alternatives used in point-to-point links are patch antennas, multi-element Yagi antennas and parabolic dish antennas.

Two examples: A 2.4-GHz, 15-element Cushcraft Corp. Yagi antenna has a beam width of 30 degrees and a gain of 14 dBi, while a 3-foot Cushcraft parabolic antenna has a beam width of 10 degrees and a gain of approximately 24 dBi.

Now the key measures are in place. You start with output power, say 20 dBm, add antenna gain and then subtract loss from cables and free space. If the resulting number still exceeds the minimum for receive sensitivity, the signal gets through. To provide some margin for error, installers will typically define a fade margin of perhaps 20 dB.

While point-to-point wireless systems require line of sight between antennas, getting it is not quite as simple as aiming the antennas by using a pair of binoculars. In some cases, you might have direct visual line of sight, but not true RF line of sight. Likewise, minor foliage might inhibit visual line of sight, but under some circumstances, a point-to-point wireless link may be able to tolerate leaves and tree branches.

Playing it Safe

Security is always a significant concern with wireless systems, but since most well-designed point-to-point links are highly directional, intercepting signals is difficult. Nonetheless, most vendors include hardware-based encryption features that do not hurt system performance. And because you are concerned about securing a single point-to-point link, you can employ simple shared keys.

Many enterprises struggle to decide whether to contract with a wireless integrator/installer or do it themselves. The main factors to consider in making this decision are the complexity of the link and your risk tolerance. Connecting two buildings separated by a public thoroughfare at 10 Mbps is a lot easier than connecting two sites separated by 20 miles at 100 Mbps. Experienced radio system installers not only have a reliable gut feel for what will work but also understand some of the subtle complexities of RF link engineering, including antenna polarization, Fresnel zones and multipath interference, and they have the tools to measure RF signals, including potential interference. They also have a good understanding of government regulations, including those related to acceptable EIRP (effective isotropic radiated power), a measure of total system gain in specific frequency bands.

In addition, professional installers have experience with antenna and lightning arrestor installation.

In selecting specific products for installation, consider your throughput and reliability requirements and the RF characteristics of available systems. Think about output power, receiver sensitivity, management and security capabilities, mean time between component failure, and cost. Many fixed wireless systems are adapted from WLAN technology, while others are designed specifically for point-to-point applications. The latter generally offer greater reliability to meet the standards imposed on them by service providers, which are major customers. You can spend $10,000 for a system or 10 times that amount, but except for typically modest ongoing maintenance costs for hardware and software, monthly service charges paid to your local telco will be a thing of the past.

Dave Molta is a senior technology editor of Network Computing. He is also an assistant professor in the School of Information Studies at Syracuse University and director of the Center for Emerging Network Technologies. Molta's experience includes 15 years in IT and network management. Send your comments on this articles to him at [email protected].All antennas radiate energy in three dimensions. However, while the theoretical isotropic antenna radiates along the X, Y and Z axes equally, real antennas don't behave that way. Some antennas radiate energy in a symmetrical pattern, while others are asymmetrical in nature.

The diagrams here were originally presented in an article by Trevor Marshall that appeared on Byte.com, Antennas Enhance WLAN Security. They provide 2-D and 3-D representations of antennas commonly used for WLAN applications. The same general principles apply to point-to-point antennas.

The first diagram shows the 2-D and 3-D patterns for a typical dipole antenna. This antenna provides gain by radiating more energy along the X axis than the Z axis. This is more precisely shown in the 2-D diagrams, where the azimuth view shows the pattern as viewed from above, and the elevation pattern shows the pattern as viewed from the side. The gain of this antenna is approximately 2.1 dBi.

The second diagram represents the pattern of a biquad antenna. This system radiates very little energy along the Z axis, instead concentrating it in a single direction along the X axis. The gain of this antenna is 11.3 dBi. The biquad antenna is designed to provide directional coverage within a building, so its beam width is wide. For point-to-point applications, narrower beam widths are highly desirable because the goal is to focus energy on a specific, distant point. Thus, for Yagi and Parabolic antennas, the signal along the X axis would be longer and narrower.

Most antenna manufacturers provide 2-D representations of their antenna patterns, so it's good to get accustomed to reading them in that manner. However, software is available to convert those 2-D patterns to 3-D (see Antennavis 0.2 and NSMA Antenna Pattern Plotter for Windows).

An interesting tutorial on antennas can be found at the AeroComm Antenna Tutorial.