 |
 |
|
||
|
|
Concepts to Understand about Wireless Networks |
| CCNA candidates need to have a basic understanding of wireless and this article by Ken Chipps, excerpted from a CertificationZone Tutorial titled, "How to Implement Wireless Networks" provides a lot of the needed information. |
A radio uses radio waves to send and receive information. These radio waves are part of the electromagnetic spectrum. Electromagnetic radiation fields consist of two planes. It is necessary to understand what a plane is in relation to a radio wave front because it has an impact on the ability of two ends of a wireless radio link to communicate. These planes are:
E Field (Electric Field) - Exists in a plane parallel to the antenna
H Field (Magnetic Field) - Exists in a plane perpendicular to the antenna
In other words, the E field lines up with the antenna. Using a dipole antenna as an example of this the E field aligns this way.
Related to the two planes exhibited by an antenna is the polarization of the signal. Radio frequency signals are polarized. The polarization aligns with the E field. If the electric field lines are parallel with the surface, then the polarization is horizontal. When those electrical field lines are perpendicular to the surface, the polarization is vertical. The antenna type and alignment determine the polarization of the radio wave. For maximum signal strength, the antennas at both ends of the transmission must use the same polarization. This is difficult to do in a wireless local area network. However, it must be done for the longer length connection used in a point-to-point campus area network.
Even if the two ends of a wireless network have their polarization perfectly aligned, signal strength is still lost due to several factors that affect the flow from the transmitter to the receiver.
All radio links suffer from loss. This loss can be due to the parts used or the environment through which the signal travels. Loss cannot be eliminated, just controlled.
Amplifiers produce gain in a radio. In general, do not change or add an amplifier to a radio system you buy. There are arguments for and against this practice. I do not use add-on amplifiers. If more gain is required, the gain of the antennas -- which is different from amplifier-produced gain -- should be changed instead.
To measure loss or gain, a common unit is required. The decibel is the unit used. In the systems we are discussing it takes several forms.
Because the decibel is a ratio between two power values, such as input and output power, another measure is needed to express power in terms of a fixed reference point. This is the dBm. This uses 1 mW (milliwatt) as the standard. 1 mW = 0 dBm.
The dBi refers to the gain of an antenna in relation to a theoretical isotropic radiator. The isotropic radiator radiates in a perfect sphere around the antenna. The radiation pattern looks like a soccer ball or basketball with the antenna element in the center. When gain in dBi is produced, it is through redirection, or focusing, of the antenna's output. This behavior is similar to a light bulb that has no reflector, outputting its light in all directions, compared to a flashlight's reflector that focuses the beam in a particular direction. The light bulb maintains its output level, but it is focused, or concentrated, in a particular direction. Antenna gain is always redirection. Only amplifiers increase the real gain in a system. Think of this perfect sphere now as a balloon. If one side of the balloon is pulled out in a single direction, the size of the balloon is not changed. It is just redirected. This is what happens when a real antenna replaces the theoretical pinpoint antenna in the middle of the radiation sphere. The different shapes and sizes of the real antenna always redirect or distort this perfect spherical radiation pattern. Therefore, any real antenna has some gain, but only in relation to the perfect radiation sphere of an isotropic radiator.
The first of the environmental factors that produce loss is noise. Noise consists of all undesired radio signals, whether produced by humans or naturally occurring. Noise makes the reception of useful information difficult. The radio signal's strength is of little use if the noise power is greater than the received signal power. This is why the signal to noise ratio is important. Increasing receiver amplification cannot improve the signal to noise ratio since both signal and noise will be amplified equally and the ratio will remain the same.
Naturally occurring noise has two main sources: atmospheric noise, such as thunderstorms, from 0 to 5 MHz; and galactic noise, such as stars, at higher frequencies. Both of these sources generate sharp pulses of electromagnetic energy over all frequencies. The pulses are propagated according to the same laws as the desirable signals being generated by the radio equipment. The receiving system must accept them along with the desired signal.
The noise produced by human beings is part of modern life. It is generated almost anywhere that there is electrical activity, such as automobile ignition systems, power lines, motors, arc welders, fluorescent lights, and so on. Each occurrence is small, but there are so many that together they can completely hide a weak signal in an urban area that would be above the natural noise in a less populated area. This is also one form of a DoS attack that may be performed on a wireless network. An attacker can intentionally transmit garbage in the same frequency range as the wireless network and, providing they do so at sufficient power, will render the WLAN unusable.
The measure of the effect of the noise in the environment is the signal-to-noise ratio (SNR). If the signal is more powerful than the noise, then reception is possible. The signal to noise ratio is the difference between the signal and the noise dBm values. To compute the SNR, given values in dBm, use this formula.
SNR (in dB) = Signal_strengh - Noise_strength (in dBm)For example, for a signal value of -45 dBm and a noise reading of -92 dBm, subtract -92 from -45. The result is a SNR of 47 dB. Some site survey programs, such as the Cisco Aironet Client Utility may show a slightly different value.
SNR can also be calulated using the ratio of the signal and noise power levels (in Watts, for instance):
SNR = 10 * Log10(Signal_Power / Noise_Power)
Many site survey programs show the computed SNR along with the signal and noise values.
Effective Isotropically Radiated Power (EIRP) is the power actually radiated by the radio system. It is the product of the power supplied to the antenna from the radio and the gain of the antenna. Governmental authorities regulate this power level.
For example
Keep in mind that the radio frequency environment is dynamic. In this way, the radio wave environment is similar to the weather. Just like the weather, we know quite a bit about how large systems operate. What we cannot predict is the microclimate for a locality. This is also true of radio frequency performance. The propagation of radio waves is well understood. Exactly how these waves will or will not penetrate a particular building is not. This is the main reason for the fade margin or fudge factor. This fade margin is used below when the site survey procedure calls for a line to be drawn connecting the 20 dB SNR points identified during the survey.
As radio waves move along their journey from here to there, many things act them upon. Most of these things are not good. Everything a radio frequency (RF) signal encounters on its journey has an effect on the signal. The effect is usually to make the signal smaller or to change its direction in some way.
The first of these impediments is free space path loss. This type of loss occurs regardless of whether the signal is transmitted inside or outside a building. Free space loss is the widening out of the signal as it moves away from the antenna. The result is lower signal strength at the receiving end of the link. Propagation loss increases with respect to both distance and frequency. In other words, higher frequency signals lose more than lower frequency signals, because the short wavelengths of the higher frequencies cannot bend around objects as well as longer wavelengths. The practical effect of this is that short wavelengths are line of sight links. This is not a consideration for the size of networks discussed here: local and campus area. However, it is a consideration for a longer link.
Free space loss is only one of the losses suffered by the signal as it goes from here to there. Absorption is another loss. It is caused by things that the signal runs into. For the wavelengths and networks of interest to us, the amount of loss experienced by a radio wave from absorption depends on the materials the wave encounters on its journey.
Inside an office building, the absorption is from the building materials, furniture, and so forth that the signal encounters. As human beings are basically large bags of water, it is assumed -- and has been reported in the trade press -- that too many humans in a space will cause signal levels to drop. My informal testing in classrooms does not confirm these anecdotal reports. However, a fairly rigorous study by Intel does show loss due to water in the path of the signal in a residential environment. This water was not in humans, but jugs of water placed between the two radios. In a warehouse, the impact of water on the signal coverage is more pronounced. For example, a beverage packaging operation I examined sees the coverage area of each of the access points in their product warehouse reduced by up to 60 percent when the warehouse is full versus when almost empty. However, it is difficult in this instance to separate the effect of the metal beverage cans from the liquid they contain. More work is needed on this topic.
In the types of systems used to create campus area connections, the main outside absorption problem is vegetation. There are no firm numbers for this problem, but some general statements can be made. The absorption is due to the water content of the vegetation and the frequency of the signal. For full foliage trees in the Northern Hemisphere, research suggests these values for the absorption effect of vegetation:
| Frequency | Absorption | |
| dB per meter of foliage | dB per tree | |
| 870 MHz | 0.2 to 1.3 | 11 |
| 1.6 GHz | 0.5 to 1.3 | 11 |
| 5 GHz | 1.2 to 2 | 20 |
Studies suggest that the wood part of the tree is the major factor in tree-related attenuation at frequencies from 870 MHz to 4 GHz. Leaves add up to 35% additional attenuation at 870 MHz, plus an additional 15% at 1.6 GHz. At 20 GHz, the wood and leaves are both important. An International Telecommunication Union (ITU) study on this subject states that the attenuation caused by vegetation varies widely due to the irregular shape of vegetation, as well as the wide range of sizes, shapes, densities, and water content of various species.
Reflection is a change in direction of the signal caused by something the signal cannot penetrate. The amount of reflection depends on the wavelength, the material the object is made of, and the angle at which the signal strikes the object. Reflection occurs when the object has a very large dimension compared to the wavelength. As most of the wavelengths used in wireless systems are very short in comparison to the objects they encounter, many things in the environment cause reflection. If the material does not absorb the entire signal, some must bounce off (i.e., be reflected). A smooth metal surface with good electrical conductivity exhibits severe reflection. Reflection appears as multipath.
Refraction is the bending of a wave as it passes through an object. It is not reflected, but mostly passes through the object. The signal that passes through goes off in a direction different from that in which it entered the object. The obstruction that causes the refraction is not always obvious. Usually both refraction and reflection occur at the same time.
Diffraction occurs when an RF wave is obstructed by a surface that has sharp edges, such as the corner of a hall or the edge of a building. The signal moves around the object and back to the other side. But a shadowed area appears behind the object.
Scattering occurs when radio waves hit a large number of objects whose dimensions are smaller than the wavelength. In the frequencies of interest to us, common causes of this are signs in halls, foliage, and other such things found in the environment. In the wavelengths used by the unlicensed bands, many things the wave front encounters fit this definition.
These signal-reducing factors are intermingled. There is no way to separate and isolate the effects of one source as opposed to another source when a network is deployed in the real world. This is the reason for conducting an onsite site survey. One example of the combination of these effects is seen in a test recently done to examine signal distance with an access point in a hallway versus the access point in a room with both at table height. With the access point in the hallway of the building, the signal traveled 248 feet before the signal dropped to -82 dBm. With the access point in a room, the -82-dBm point was reached after 118 feet. In these 118 feet there are four sheetrock and metal stud walls, classroom furniture, and a few people.
For planning purposes, the results of the Intel study can be used. Measurements were taken in a residential townhouse of typical construction. This study produced these values for signal loss in dB.
| Obstacle | 2.4 GHz | 5 GHz |
| Wall | 10.7 | 14.9 |
| Floor | 5.5 | 7.0 |
| Water* | 3.8 | 14.2 |
*The water was 3 gallons in one-gallon jugs arranged in a triangle
In each case, signal loss was higher for the 5 GHz frequencies. Higher loss figures can be expected in an office building where metal wall studs and concrete floors are more common.
Suggested figures for estimating loss in a typical office building are:
| Obstacle | 2.4 GHz |
| Wall | 5 to 10 |
| Floor | 15 to 25 |
The figures should be somewhere in between these two studies for open plan offices.
These are all just general guidelines. In the lab, this may be science; in the field, it is all art. This discussion just emphasizes the need for the onsite site survey.
Since the early 1900s, the use of the radio frequency spectrum has come under more and more regulation by governmental bodies.
As might be expected when dealing with governmental bodies and their rules, there is a multitude of regulations, including:
International
ITU-R - International Telecommunication Union, Radiocommunication Sector
Europe
CEPT - European Conference of Postal and Telecommunications Administrations
ETSI - European Telecommunications Standards Institute
ECC - Electronic Communications Committee
ERO - European Radiocommunications Office
Western Hemisphere
CITEL - Inter-American Telecommunication Commission
Canada
Industry Canada
Spectrum Management and Telecommunications Sector
United States of America
FCC - Federal Communications Commission
Asia-Pacific
APT - Asia-Pacific Telecommunity
Japan
ARIB - Association of Radio Industries and Businesses
Australia
ACA - Australian Communications Authority
New Zealand
Ministry of Commerce
Most devices and frequencies operate under either:
one or both of the regulations created by the FCC in the United States of America under Parts 15 and 101 of Title 47 of the Code of Federal Regulations
ETSI EN 300 328-1, ETSI EN 300 328-2, and ETSI EN 301 893 from the ETSI in Europe.
For the unlicensed bands, most countries use the requirements of either the FCC Part 15 or the ETSI EN 300 328.
Harmony among the worldwide rules is desirable for equipment manufacturers. It is an advantage to users as well, because wider usage generally results in lower costs. For the unlicensed 2.4 GHz frequency range, most of the worldwide restrictions have been lifted, except for a few channel restrictions and limitations to indoor use only. France, Israel, parts of Latin America, Asia, and the Middle East are the remaining problem areas.
For the 5 GHz ranges, some parts are available everywhere. The differences are what parts can be used where, such as indoors or outdoors. The World Radiocommunication Conference held in 2003 harmonized and expanded the spectrum in the 5 GHz frequency range. Worldwide, these ranges track the US usage shown below. In addition, a range from 5.470 to 5.725 GHz will be added for use both inside and outside. Depending on the ultimate use of these frequencies in actual products, this should increase the available bandwidth for devices using these frequencies.
There are two key points in the regulations governing the types of networks we are discussing. First, these radio frequencies are unlicensed. This does not really mean what it seems to say. By unlicensed, the regulators mean that anyone can use these frequencies for anything as long as they conform to the rules. This means that, if a neighboring system is following the rules, but at the same time it destroys the usability of your network, you can do nothing about it. Cooperation is the key to getting along in this unlicensed space. Using a licensed frequency for this type of network is not practical. The cost for the license and the equipment is too high.
The second aspect is controlling interference in this unlicensed space. The regulations specify the power levels that can be used, the way the parts are connected to each other, and the required modulation techniques.
When installing a wireless network using these frequencies, you cannot just do anything you want to. You are limited to approved radio systems. This does not mean you must use only Cisco antennas with Cisco access points. It does mean that the supplier of the antenna must have an approval from the regulatory body for that antenna and radio combination.
There are a limited number of available frequency bands for use in local and campus area wireless networks. All of these frequencies are unlicensed. This means that the maximum legal power is limited and interference from other systems must be accepted.
In many parts of the world, the 900 MHz frequencies are license-free in the range from 902 - 928 MHz. This total bandwidth is 26 MHz. The nominal wavelength is about 325 mm. This frequency is not currently used in the types of networks we are discussing. It was used in some early LAN equipment. It is regaining popularity in wireless Internet service provider deployments for its ability to penetrate vegetation. It is mentioned here because point-to-point campus area links may start to use it. The main problem is a lack of bandwidth. Thus, data rates are slow.
The 2.4 GHz frequency range is license-free worldwide for the most part, although the channel details differ. It ranges from 2.4000 - 2.5000 GHz in the FCC scheme in the United States and from 2.4000 to 2.4835 GHZ as defined by the Institute of Electrical and Electronics Engineers (IEEE), which creates standards for the use of these frequencies. The 2.4 GHz band provides 83.5 MHz of usable bandwidth. These signals are around 125 mm long. This is a fairly long-range solution, but it requires line of sight, because it does not pass through obstructions well. Water attenuation is the major problem, especially outdoors. The attenuation from trees is approximately .5 dB per meter of canopy. With a tree with a canopy of 10 meters (~30 feet), the attenuation would be about 5 dB. Just a few trees will block the signal. In the United States, Part 15 of the Code of Federal Regulations (CFR) covers the usage of this frequency. In Europe, regulation of this frequency range is covered by EN 300 328 and EN 300 826 from the ETSI. Part 15 and EN 300 328 are similar. When used inside, the IEEE 802.11b standard is the most widely deployed system. Because this frequency range is highly utilized, interference may be high.
5 GHz systems are deployed around the world, but the allowed frequencies vary widely. For example, in the US there are four license-free subbands at 5 GHz, although two of these bands overlap each other. There is one Industrial, Scientific, and Medical (ISM) band from 5.725 to 5.850 GHz. There are three Unlicensed National Information Infrastructure (UNII) bands.
| Range (GHz) | Use | FCC Max Power (mW) | IEEE Max Power (mW) | |
| Lower | 5.150 - 5.250 | indoor only | 50 | 40 |
| Middle | 5.250 - 5.350 | indoor and outdoor | 250 | 200 |
| Upper | 5.725 - 5.825 | outdoor only | 1000 | 800 |
The ISM band is 125 MHz wide. Every UNII band is 100 MHz wide. The 5 GHz range wavelength is approximately 54 mm. An important point for future development is that each 5 GHz subband is wider than the entire 2.4 GHz band. It is possible to build 5-GHz wireless equipment that provides more bandwidth and more throughput than equipment for any other unlicensed band. Part 15 covers the 5 GHz band in the US. In Europe, this group of frequencies is generally defined under EN 300 440 and EN 300 683, which cover all frequencies from 1 to 40 GHz.
The attenuation from trees at 5 GHz is about 1.2 dB per meter.
|