Steps in the Site Survey Process
RF engineering for specific sites is not an exact science, and
therefore neither is the practice of doing site surveys. There are many, many
ways to perform a site survey, and of all these different methods many are just
fine. Because one engineer prefers method one but another engineer prefers
method two does not necessarily make one method better or more correct than the
other. However, both methods should provide similar resultsthat is, adequate
coverage and performance.
The general steps (some of which have been discussed in various
chapters in this book already) that should be part of a site survey include the
following:
|
Step 1. |
Obtain a floor plan or facility blueprint. |
|
Step 2. |
Visually inspect the floor plan. |
|
Step 3. |
Identify user areas on the floor plan. |
|
Step 4. |
Identify potential problem areas on
the floor plan. |
|
Step 5. |
Identify AP locations and antenna types. |
|
Step 6. |
|
These are general steps, and each one builds on the other to
complete a quality survey. Skipping any of these steps can make for problems
during the installation or the actual implementation of the WLAN. Several topics
have been discussed previously, and in this chapter details of the remaining
topics are covered.
Obtain a Floor Plan or Facility Blueprint
You have learned about obtaining a floor plan or facility
blueprint in Chapter 6, "Preparing
for a Site Survey," and more details are included in Chapter 13, "Preparing the Proper Documentation," so
the discussion of such is kept brief here.
Many survey tools, such as the AirMagnet Surveyor and Cisco
Wireless LAN Solution Engine (WLSE), enable you
to import many types of standard graphic formats such as JPEG and BMP files.
This makes importing easy. If a drawing is not available, you should generate
one, even if it is somewhat simplistic.
Make this drawing available in a printed version so that the
engineer can easily make notes throughout the survey. Without this capability,
it will be difficult to capture all the necessary information needed.
Inspect the Facility
When possible, personally visit the site before the survey
begins. On-site visits enable you to understand the working conditions under
which the WLAN will be used and the conditions under which you will perform the
survey. A site visit will also help you determine which types of antennas might
be used in a facility, determine whether any particular pieces of equipment
might be needed for the survey, and give you a chance to inspect the overall
site for unusual issues or concerns that could inhibit the completion of the
survey.
This is the time that safety issues should be analyzed, and
determinations made as to whether safety equipment or special clothing is
required. It also gives you the chance to meet the on-site personnel and learn
of any issues or concerns they have regarding your survey work.
Identify User Areas on the Diagram
User density should have already been
specified in the initial WLAN design. Understanding the location of the
userswhere they gather in numbers and where the WLAN is not neededis critical.
These parameters should be well defined on the facility diagram. Make note of
areas such as lunch rooms, conference areas, informal gathering areas, and even
outdoor areas where users gather in numbers.
Another portion of this documentation task is to identify areas
that have the common characteristics for physical user areas and density. For
example, cubical areas tend to be similar in physical size and distribution in
the same facility, making these areas very similar for user density. Similarly,
conference rooms, meeting rooms, and so on, if similar in physical size, can be
considered as like areas. Note any areas where the contents, applications to be
used, number of users, and geographic location of users are similar. This helps
when performing the user-density testing and is vital for using assisted site
survey tools. The different types of areas might have different requirements for
cell size and bandwidth.
Another task here is to identify areas where users require a
constant connection while roaming from one point to another. In most cases,
laptops are not used when physically in motion and represent a very small issue
here. However, voice over IP (VoIP) phones, PDAs,
bar code scanners, and other similar devices are commonly used while moving
through some facilities, and you need to consider their requirement for constant
connection.
Identify Potential Problems Areas on the Diagram
One of the first steps in a survey is a walk through the
facility, which is needed to search for potential problem areas. From the
diagram itself, you often can identify building-structure items such as
elevators shafts, stairwells, and fire doors. These items tend to cause RF
shadow areas and dead spots. If coverage is needed in these areas (as may be
desired for seamless roaming with voice, for instance), take special care to
place APs strategically for coverage.
Also note other wireless equipment or devices in the facility
that might cause interference. This information can help you reduce problems
that might occur if you place APs too close to these devices or systems (or
prevent placement on the opposite side or a wall, hidden to the eye, but not the
RF).
It is also a good idea to note items that might have RF
reflections, which in turn might cause multipath signals, such as large metal
doors, metal cabinets, or, as commonly found in many warehouses, secure areas
that use a metal meshing as the walls. (See Figure 11-1.) This notation helps explain why an AP is
placed in a particular location, or why a particular type of antenna is
used.
Tip
While noting potential problems, be aware of things that can
change in the environment. For instance, large metal doors might be open at the
time you perform the survey, but might be closed at other times. Normally you
want to perform a survey with all doors closed, even if they are usually kept
open. In general, it is better to have a little extra overlapping coverage via
open doors than to have closed doors blocking the RF signal and causing dead
spots.
Identify AP Locations and Antenna Types
Identifying AP
locations and the antennas to be used is the heart of any site survey project.
Exactly how you identify these particulars will vary from one type of survey to
another and from one site survey tool to another, but the overall results (that
is, coverage and user performance) should be similar.
As you perform a survey, you need to have a good understanding
of RF issues in the site, such as RF signal attenuation and how to identify RF
interference. This knowledge will assist in identifying problem areas in the
site. To determine where to place the AP and what antenna to use, a walkabout is
performed and RF measurements are taken. These measurements are used to
determine the RF coverage, identifying the edges of the APs' cell boundaries.
This step also helps to identify the overlap needed between cells to ensure
continuous RF coverage with no dead spots. Each of these topics is discussed in
detail in the following section (and the differences between manual and
automated/assisted survey methods are noted).
RF Issues in the Site
As discussed in Chapter
2, "Understanding RF Fundamentals," one of the first issues to be understood
regarding WLAN site surveys is that distance between the AP and client affects
WLAN bandwidth and therefore capacity. Unfortunately, when deploying any radio
system, including WLANs, the laws of physics apply. RF signals propagating
through the air are subject to attenuation, losing signal strength while
encountering obstacles, both natural (including the atmosphere) and manmade.
To effectively deploy a WLAN, technicians must understand the
causes of RF attenuation and what applicable countermeasures are available. This
knowledge is extremely important to the site survey engineer as well.
As a rule, as the distance between a client and an AP
increases, the signal strength decreases. And at some point the bandwidth has to
decrease to maintain the connection. The actual attenuation will vary widely,
and without testing and verification it is impossible to determine exactly the
overall effect of all the objects in most sites.
As described in Chapter
2, as the frequency increases, the amount of attenuation produced by the
atmosphere increases, reducing range. Unlike outdoor line-of-sight applications
based on straightforward path-loss calculations, attenuation for indoor systems
is much more difficult to calculate. The main reasons for this difficulty are
the multipath signals that occur in most indoor sites and the different
attenuation effects created by the various materials found indoors.
The algorithms used to estimate path
loss are very complex and are used in the theoretical site survey
tools. These algorithms differ from indoor to outdoor, and the attenuation of RF
signals also vary. In Figure 11-2, notice
that at distances up to 50 feet attenuation is very similar. Beyond this
distance, however, the path loss indoors increases much faster. Even so, you can
estimate that indoors approximately 100 dB of attenuation occurs over distances
of 150 to 200 feet for 2.4-GHz signals. Remember, however, that attenuation is
not linear and it increases exponentially as range increases.
Typical obstacles found indoors, such as walls, doors, and
office furnishings, offer fairly consistent levels of attenuation. Some standard
items can be estimated for attenuation, as shown in Table 11-1.
Table 11-1. Attenuation for Standard Building
Obstacles
|
Building Obstacle |
Attenuation
Level |
|
Drywall |
3 dB |
|
Plaster and lathe wall |
4 dB |
|
Cement-block wall |
4 dB |
|
Typical glass window (nonmetallic tint) |
2 dB |
|
Steel-reinforced preformed concrete wall |
9 dB |
Some obstacles in the site might offer such a high level of RF
attenuation that little or no RF penetrates it. Such a scenario usually results
in an area with a high concentration of metallic content, such as steel floor
pans, steel reinforced walls, metal mesh behind the stucco walls, elevators shafts, and so on. However, a high
concentration of paper, cardboard, or other materials that contain a high level
of moisture can causes serious problems for RF penetration, especially at 2.4
GHz. Even some types of tinted glass or energy-efficient glass can cause a high
level of attenuation or RF shadows.
When performing a walkabout survey, note areas where the signal
drops off rapidly. This generally indicates some type of RF shadow effect and
might require a different placement of the AP and antenna so that the area has
adequate RF coverage. In some cases, you might need to actually add an AP if the
area is totally blocked (such as an x-ray room, metal-walled freezer area, or
steel-reinforced section of a building).
Identifying these areas is critical for full coverage and for
proper roaming with devices such as wireless phones. Site surveys techniques
that do not use a client walkabout to assist in the survey often do not identify
areas that will exhibit poor roaming. A totally automated survey just cannot
guarantee there are no RF shadows and dead spots associated with shadow
effects.
One possible problem mentioned earlier is fire or hazard doors.
These are doors that must remain open during normal business hours, but close
automatically to restrict the spread of a fire or other hazard. Although many
sites do not anticipate using the WLAN during such an emergency, some sites
might want to use voice or even PDAs to assess the hazard problem. Surveying
with the doors open will likely lead to dead spots when the doors are closed.
Obtain permission (and use signage to indicate work is in progress) to close the
doors if the survey is being performed during business hours.
While moving throughout the facility during the survey process,
you will find that the signal levels vary in strength. However, the amount of
variation should not be pronounced over small movements. Because RF is an analog
signal with many influences on its strength and propagation, the overall level
will be somewhat linear over small movements, unless acted upon by an obstacle.
Dramatic variations that occur when moving only short distances (inches and
feet) can indicate a high level of multipath interference. For such areas, there
are several possible solutions:
-
Use diversity antennas (recommended in any case)
-
Verify diversity is set to on in the AP
-
Move the AP (actually the antennas) farther away from any metal
structures
-
Use a more directional antenna
Detecting Interference
As discussed in Chapter
6, the pre-site survey form should have a place to identify any known RF
systems that are used on the site. However, not all sites will have a single
document or even someone who knows all of the RF equipment that is in use. Most
enterprise WLAN equipment has the capability to look for other WLAN devices
(usually referred to as rogue AP detection), and
some can even report other interfering signals. The drawback to such features is that the APs for the new WLAN
need to be installed first. If the installer is not aware of the other WLANs, he
might install an AP in very close proximity to an interfering device. This will
in turn require a relocation of the AP, possibly affecting other AP locations as
well.
It is vital to identify all other possible interference in the
site before starting any RF survey work (through a walkabout survey, for
example). Even if the survey to be used is an automated survey, it is a vital
part of an installation to first look for and identify any interfering
signals.
There are a number of methods to identify potential
interference. The most accurate is to use a spectrum analyzer. Spectrum
analyzers enable you to view the entire spectrum, looking for signals that might
not only be within the frequency range of the intended WLAN system, but could be
near or at a frequency that could cause interference.
|
Interference can come in many flavors. You have already learned
about the problems with microwave ovens at 2.4 GHz, and wireless phones at
either 2.4 or 5 GHz. Many other devices can cause interference, too; and
although some of these will be in the same band as the WLAN, some might not
be.
Many vendors now offer wireless cameras, wireless security
systems, wireless theft-detection systems, and even wireless projectors. Because
the 2.4- and 5-GHz bands are unlicensed, these devices often fall into these
bands. Efforts should be taken to identify and document the frequency and power
output as well as the level of interference that these devices create for the
WLAN. When evaluating the interference, be certain that the device is running at
full RF capacity.
There are also devices that might be in a totally different
part of the RF spectrum but can cause RF interference. One fundamental
characteristic of an RF signal is known as the harmonic
of the signal. When a signal is generated (and transmitted in the case of
an RF signal), harmonics are also created. These harmonics are multiples of the
desired signal, and although at a much lower level (both by design and by
regulation limits) they can at times cause interference. The odd harmonics or
multipliers are the strongest, with most often the third harmonic (three times the
transmitted frequency) being the problem. As the harmonics climb in frequency,
the signal level of the harmonic usually decreases very quickly.
Consider, for example, an installed 802.11g WLAN that is using
channel 6 (2442 MHz). The site is a small-city office building where the police
department and government offices are combined. Located at the same site is the
citywide communication tower, which is attached to the building for support (and
therefore very close in proximity to the WLAN in the building). Also located on
that same tower are other wireless services that lease space from the city. If
that tower happens to have an 814-MHz transmitter on it, there is the
possibility of interference to channel 6 (814 MHz * 3 = 2442 MHz = 802.11g
channel 6). A WLAN system with rogue AP detection would never find this problem
because it is not an 802.11 signal. Therefore, a spectrum analyzer would likely
be needed. However, a spectrum analyzer in the hands of an inexperienced user
might not reveal it either, because the interference would be there only when
the 814-MHz transmitter is actively transmitting RF, and might be easily
overlooked if the analyzer is not set up
properly. |
It is vital to the quality of interference detection to become
proficient with a spectrum analyzer. To locate any possible interference from
some non-802.11 transmitter (see the "Interference from Non-802.11 Equipment" sidebar), use a
higher-gain antenna on the analyzer, a peak hold function to capture any signals
that are on line for a short period of time, and proper resolution and video
bandwidth settings.
Existing WLAN devices represent another common source of
interference. If the existing device uses a separate band (900 MHz, for
instance), then this should not cause an issue. However, it is still recommended
to keep some minimum distance between any two RF devices (minimum of 3 feet, or
about 1 meter) even if they are on different bands.
When installing a system in the same facility that has
competing RF on the same band, exercise extreme caution during the installation
to keep interaction to a minimum. For example, when adding an 802.11b or 802.11g
system to a site that has an existing frequency-hopping (FH) system, maintain a minimum of 10
feet (3 meters) between the 802.11b or 802.11g and the FH system RF
components.
Another common issue with regard to interference is the rogue
AP (that is, the AP that some employee has brought in and put into the network
without the consent of the IT staff). This type of AP can cause several issues,
with the number one being security (because rogue APs typically do not conform
to the IT security requirements).
The second issue is interference with the properly installed
WLAN. If not identified before the walkabout portion of the survey begins, it
can cause missed packets and higher noise-floor readings, which in turn might
trigger the need for another AP in that location. These devices should be
"sought and destroyed" before starting a survey.
Some WLANs systems offer rogue AP detection utilities, but
require the WLAN to be fully installed and operational before they can be used.
Therefore these utilities are more for maintaining a WLAN and identifying rogue
APs in an operation WLAN and not for use as part of a site survey.
You might feel like there is far too much to think about
regarding interference, but that is not necessarily the case. Although 2.4 GHz
does have many more possibilities of interference (because of more devices on
the market), in reality you will have few interference problems with a WLAN when
surveyed and installed correctly.
The Walkabout Test
The walkabout is one step in
performing any site survey and completing any WLAN design. Although the process
is nothing more than actually walking through a facility, taking measurements,
and verifying coverage levels, it is a vital step that must not be overlooked,
and it must be done logically. During the walkabout, you verify the signal
attenuation of objects, define cell boundaries, identify noise-floor problems,
and verify communication between a client and an AP at the appropriate data
rate, all of which help to determine where to place an AP and what type of
antenna to use. As stressed throughout this book, skipping the walkabout test
will result in a WLAN with one of two problems: dead spots or a highly
overengineered, and therefore highly expensive, WLAN. A proper walkabout mixed
with other survey techniques and appropriate design is the only way to guarantee
the best performance, efficiency, and economy. But how is this vital step
completed? Take a look at the steps to identify the boundaries of a wireless
cell, as well as to verify the proper overlap of coverage between cells.
Defining the Cell Boundaries
When you start to do either a user-density test (discussed
later in the section "Performing a User-Density Test") or a
manual survey, you need definitions for the cell boundariesthat is, what
constitutes the edge or limitation of the cell. To establish the cell
boundaries, you first need to define the following parameters:
-
Packet
size As discussed in Chapter
10, "Using Site Survey Tools," one of the parameters that needs to be set
correctly is the packet size. Packet size is dependent upon the applications
that will be used in the site. This should be set to the largest packet that
will be used. If the system will be used for standard Ethernet access, the
packets size should be set to 1400 bytes (or as high as 1518 depending on the
limitations of the site survey tools).
-
Data
rate The AP should be set to the minimum data rate permitted in the
design.
-
Transmitter
power TX power needs to be set to either the maximum transmitter level of
AP (if you are using the same AP model for the survey as for the installation)
or to the maximum level of the AP intended for deployment.
So now that you have set the appropriate parameters in the AP
for performing the RF tests, it is time to place the AP in a location in the
site and take some measurements. Before proceeding, however, you need to define
the measurements and RF test results that will determine the cell
boundaries.
As mentioned in Chapter
10, you use three major items to determine whether the signal is adequate
for proper WLAN performance:
-
Signal strength
-
Noise level
-
Packet retry counts
Together these three can provide not only a good indication of
cell boundaries, but also assist the survey engineer to understand why
communication issues exist at certain points.
As the client is moving away from an AP, the signal level will
be getting lower overall. There will be some fluctuations because of multipath
signals, but in overall scope the signal level should gradually decrease. To
determine the edge of the coverage for a data network, refer to the values
listed in Table 11-2.
Table 11-2. Cell Boundary Recommendations for 2.4-GHz
Data Surveys
|
Data Rate |
Absolute Minimum RX
Threshold |
Recommended Minimum RX
Threshold |
Absolute Minimum RX Signal to Noise
(S/N) |
Recommended Signal to Noise
(S/N) |
|
54 |
71 |
61 |
25 |
35 |
|
36 |
73 |
63 |
18 |
28 |
|
24 |
77 |
67 |
12 |
22 |
|
12/11 |
82 |
72 |
10 |
20 |
|
6/5.5 |
89 |
79 |
8 |
18 |
|
2 |
91 |
81 |
6 |
16 |
|
1 |
94 |
84 |
4 |
14 |
|
Packet Retry Rate |
Less Than 10% Packet Loss |
The table includes four columns of RF signal-level values. The
absolute minimum RX threshold (receiver sensitivity) indicates
the absolute minimum performance of the receiver at the given data rate. This
particular table contains the values for the Cisco Aironet 802.11a/b/g combo
card. These values should be set according to the equipment that will be used
(the worst-case device). Notice that the minimum recommended RX threshold is 10
dB greater (less negative) than the RX threshold. This provides a 10-dB margin
for fluctuations produced by multipath, body movement, body shadows, and so
on.
As you understand by now, signal strength alone is not adequate
to determine coverage. The table also defines the absolute minimum S/N values
for the device to receive and decode a signal properly. Again in this case the
values are for the Cisco Aironet 802.11a/b/g combo card, but they are similar to
most other devices for the 2.4-GHz band. Next to this column is a recommended
minimum S/N value. It is also easy to see that this has the same 10-dB ratio
when compared to the minimum S/N value.
Now that you have defined the signal parameters, the final step
is to look at overall communications link quality. That is determined by the
packet performance. The minimum loss of packets should never exceed 10 percent.
Although 10 percent might sound high to engineers who have been working in a
wired network world, for RF it is normal to have a few percentage points of lost
packets. That is the nature of RF. And at 10 percent for the edges of the cell,
the retry mechanism for the data retry protocol will ensure there is no
noticeable performance impact to the user in a data environment.
If the packet loss is higher than 10 percent, and the signal
strength is also high, verify the noise floor and S/N values. A high noise floor
can cause loss of packets. If the noise floor increases, the minimum signal
level will also need to increase. Watch for large fluctuations in packet loss
and signal strength, indicating an area where multipath is very likely.
Notice that Table 11-2
is for data communications. Wireless voice is a whole different beast than data
and requires different minimum recommended values for cell boundaries. If the
WLAN system is going to be using voice, overall cell boundaries need to be a bit
stronger. Table 11-3 shows some changes
to the recommended values.
Table 11-3. Cell Boundary Recommendations for 2.4-GHz
Voice Surveys
|
Data Rate |
Absolute Minimum RX
Threshold |
Recommended Minimum RX
Threshold |
Absolute Minimum RX
S/N |
Recommended
S/N |
|
54 |
71 |
56 |
25 |
40 |
|
36 |
73 |
58 |
18 |
33 |
|
24 |
77 |
62 |
12 |
27 |
|
12/11 |
82 |
67 |
10 |
25 |
|
6/5.5 |
89 |
74 |
8 |
23 |
|
Packet Retry Rate |
Less Than 1% Packet Loss |
In Table 11-3, notice
that the margin for the recommended minimum signal level and S/N values has
increased to 15 dB. This is due to the nature of voice, and the critical
necessity for maximum performance in packet transfer. Missed packets in voice
are immediately noticeable to the user's ear. Therefore, not only has the
minimum recommended value been increased, the maximum packet loss has been
reduced to a much lower limit.
Finally, notice that the data rates below 5.5 Mbps are missing.
Although it is recommended in most voice applications to maintain an 11-Mbps or
higher data rate, a few vendors of wireless voice products do suggest allowable
usage of data rates as low as 5.5 Mbps, but nothing lower.
Figure 11-3 shows the
defined cell boundaries.
For different devices or different bands, these signal level
and S/N values will change. It is important to understand the devices' technical
specifications so the criteria can be defined. Using the same Cisco Aironet
802.15 a/g combo card, and looking at the 5-GHz performance as shown in Table 11-4, it is easy to see the
differences from the 2.4-GHz band shown in Table 11-3. Although the overall margins have been kept
consistent (for the data networks), the overall levels have changed because of
differences in the radio capabilities.
Table 11-4. Cell Boundary Recommendations for 5-GHz Data
Surveys
|
Data Rate |
Absolute Minimum RX
Threshold |
Recommended Minimum
RT |
Absolute Minimum RX
S/N |
Recommended
S/N |
|
54 |
68 |
58 |
20 |
30 |
|
36 |
73 |
63 |
14 |
24 |
|
24 |
77 |
67 |
12 |
22 |
|
12 |
82 |
72 |
7 |
17 |
|
6 |
85 |
75 |
5 |
15 |
| |
|
|
|
|
Using the parameters defined in Tables 11-2, 11-3, and 11-4,
you can define or verify the boundaries of a cell. In defining a cell, however,
you must determine where one cell ends and another begins. To put it another
way, how much overlapping coverage should occur between cells, and how do you
verify that? The next section addresses these issues.
Overlapping Cell Coverage
Just as important
as cell boundaries is the concept of overlapping cell coverage. Excessive
overlap of coverage can result in some channel interference, unnecessary
AP-to-AP roaming (by client devices that have a limited roaming algorithm), and
added expense because of more APs being required.
A typical overlap in coverage is set to about 10 percent to 15
percent of the overall cell coverage area. Some engineers try to place a minimum
signal level for both APs. Suggesting that the cell boundary of AP 1 is at 72
dBm (for some given data rate) and the signal level of the other AP at that
point is 57 dBm seems to indicate some amount of overlap. However, this can be
difficult to correlate to a percentage of overlap. Because signal levels vary
from site to site, the signal strength of adjacent cells is very much dependent
on the contents of the site.
Using a site map, as shown in Figure 11-4, with correct dimensions, you can define what
the 10 percent or 15 percent overlap is. When performing the final walkabout,
verify not only cell boundaries, but also that cell overlap is within the
acceptable range.
To summarize, cell boundaries for a 2.4-GHz 11-Mbps cell for
data use only (using the data for the Cisco Aironet 802.11a/b/g card) can be
determined by the following parameters:
-
At the edge of the coverage area, the lowest signal strength
should be 72 dBm or higher.
-
At the edge of the cell, the minimum S/N should be 20 dB.
-
Packet loss should be no more that 10 percent.
-
Overlapping overage should be defined by the site map, and
verification should be made that the adjacent AP can be heard with values
greater that the minimum recommended thresholds.
Also it is important to remember that the client settings used
in this cell boundary test process should match the actual network application
scenario, as well as emulate the RF performance of the worst-case client that
will be used in the WLAN system.
Performing a Manual Survey
The manual site survey is still by far the most popular and the
most accurate, but it is also the most time-consuming and work intensive (but
then again you don't get quality work without a little effort!). There are many
ways to get started with a manual survey, but one that is very common is the
"outside-in" survey method. This is where you start at the outside of the area
and work toward the center. It is logical and accurate and provides for a very
smooth workflow.
It is recommended that an analysis of the ambient RF
environment be performed. After that, set the test AP on a channel that has no
activity in the desired area.
Consider, for example, a typical retail site. In such, maximum
range is needed, because there are only a minimum number of users; user density
is not an issue. To start, an AP is placed near a corner of the facility, as
shown by reference point A in Figure
11-5. Next, using the parameters defined for cell boundaries, perform a
survey and determine where the edge of the coverage is. Mark this on your site
plan for a temporary AP location, as shown in Figure 11-5.
You might wonder why the AP is placed in the corner. Well, 75
percent of the signal is outside the facility, and that is what we do not want! However, this is not where the AP will get
installed; this is just for a starting point. Here the AP is located in the
corner along with a standard antenna (in this case, placed at ceiling level with
a 5.2-dBi omni antenna hanging down from the ceiling).
On the site map, locate the approximate center of the coverage
arc, as shown by reference point B in Figure
11-5. This will be the new location of the AP for testing. The rationale
here is that if the client can communicate from point B to the AP at point A,
they should still be able to communicate when the devices are reversed (with the
AP at point B and the client at point A). And a minimum amount of signal will
extend beyond point A outside the facility, but there will be adequate coverage
at point A. Just guessing where point B needs to be based on the site map could
result in more energy outside the facility than necessary (or desired), or not
enough signal to reach the inside corner, resulting in a dead spot.
Figure 11-6 shows the
revised location of the AP with the associated coverage area. Notice the corner
of the facility is fully covered as well.
Repeat the same
effort for points C, E, and G, as shown in Figure 11-7. After you complete this step, you might need
to fill in the center areas if there is still more coverage needed. Because the
number of users were defined to be higher in the stockroom, the two area APs (F
and H) were slid back slightly to provide adequate overlap and to provide more
signal into the storeroom (which results in fewer main store users on these APs
because less area of the main store is covered by these APs).
Knowing the
average coverage ranges, you can make an estimated guess as to where to place
the next AP. In Figure 11-8, point J is
selected and its overall coverage tested and noted. In this case, it provides
more than enough overlap to the adjacent cells. Take care to verify that it does
not overlap with enough cells so that it could interfere with some other cell on
the same channel. If this is the case, reducing the power level on this AP (or
moving to a smaller antenna) should be considered. Next place an AP at point K
and test to complete the site coverage.
Retail stores are typically one of the easiest sites to survey
because of their physical nature and contents. So what happens when things get a
bit more congested, less open, and coverage requirements and user densities
vary? That just takes a little more work. You can use the same scheme throughout
the facility. In some cases, however, it requires a little different logic. This
can be a more linear movement, starting at one point and moving across the
facility as you might do in a warehouse, or just working one section at a time
(for a health-care or education facility wing, for example).
Now look at a large do-it-yourself home-improvement warehouse.
This type of facility has tall racking that extends up to near the ceiling and
runs in long rows. Using an omni-directional antenna to cover something like
this is usually not a feasible solution. The use of some type of directional
antenna is more common and provides a fairly easy installation by placement
along the walls. As shown in Figure 11-9,
an AP with a patch antenna is placed at one
end of the building, with the energy directed down the aisle. The coverage is
then tested to see how much coverage is obtained, and exactly how well the RF
extends down the aisles.
In this particular case, notice that one AP and a patch antenna
provides about 3 to 4 rows of coverage, a little more than 50 percent of the way
down the building. Placing another AP several rows over, as shown in Figure 11-10, would provide verification of
coverage for adjacent aisles. Using this test, you could analyze that placing
APs in the locations shown in Figure
11-10 would provide coverage for the entire facility.
However, it might be worthwhile to test another alternative
before deciding on the final approach. Exchanging the patch antenna (8.5 dBi)
for a higher-gain (13.5 dBi) Yagi antenna might provide
enough range to fully reach down the aisles. This might eliminate the need to
place APs at the back of the facility. Because the higher-gain antennas have a
narrower beam width, however, they will likely not cover as many rows, requiring
more APs along the front of the facility. In the long run, if the number of APs
ends up being identical, it might be easier to install and require fewer cable
runs to have all the APs on a single wall. This, in the long run, can be less
expensive to install. Figure 11-11 shows
the coverage obtained with Yagi antennas.
For certain types of facilities (education and health care, for
example), you confront several issues. The first is the exposure of the
products, or the physical security. In most of these cases, the AP and antenna
have to be secured either out of sight or with some type of locking mechanism.
Therefore, antennas such as Yagis or omnidirectional, which hang down from the
ceiling, are typically not an option. This requires some thought as to possible
antenna types and placements.
In some schools, the determination has been made to place one
AP in every classroom. This gets quite expensive, but provides the overall best
bandwidth performance. Such installations use a very low 1-mW power setting, and
some even go beyond that and install an attenuator on every antenna, reducing
not only the transmitter power, but also the ability to hear distance clients.
Failing to reduce the overall radiated power would create far too much overlap
between APs on the same channel.
The IEEE 802.11d specification enables the client to change the
frequency of operation and the power levels, based on information received from
the AP. This is primarily done to provide a client that can roam from one
regulatory domain to another. In some cases, a few WLAN products enable the
client to match exactly the power settings of the AP.
For health-care facilities, in some areas the required
bandwidth varies drastically. For normal hallways and patient rooms, where the
network is used for patient records or perhaps bar code scanning of medicines,
bandwidth and redundancy are not critical to patient life and death. In most
cases, the WLAN is a requirement for normal operation; if there is a location
where the WLAN has failed, however, the user could revert back to the old paper
and pen method. If the WLAN goes down, it only affects the overall efficiency of
the work that is being done. However, in cardiac care
units (CCU) or intensive care units (ICU),
where the WLAN is used in the monitoring of the patients' vitals, this is very
different. The WLAN is used as a life-monitoring system, and it must be up at
all times. In these situations, there should be enough overlap of APs to provide
absolute redundancy of coverage. (See Figure
11-12.) In these cases, some overlap of the same-channel cells might be
necessary. There should be no area where a client cannot hear at least two
different APs.
Another way to ensure redundancy is to use APs that perform hot
standby. In a hot standby setup, two APs are mounted at each location. One AP is
in standby mode and monitors the other AP. If a problem occurs, the standby unit
takes over.
When trying to survey something such as a hospital wing or
office building with long hallways and identical offices, your best bet is not
the use of omnidirectional antennas. In many cases, low-gain patch antennas
might work well. In the case of the five floors depicted in Figure 11-13, the patch antenna would not radiate down the
hallway totally. Because it had a wide angle of radiation, however, it did
cover some of the floors above and below, so alternating ends of hallways
allowed enough coverage to "bleed over" so that full coverage was obtained. In
this case, a Yagi might have worked, but the aesthetics of the Yagi were not
conducive to the facility.
While surveying, you need to consider another dimension:
vertical. RF is three-dimensional in that it radiates in all directions.
Facilities that are more than a single floor need special attention. In these
situations, it is important to survey on floors above and below to verify where
coverage comes into play. You can reduce undesired floor-to-floor coverage by
using specific antennas. However, be sure to rotate channels of APs so that the
same channel is not used directly above or below an AP on the same channel. (See
Figure 11-14.)
Documenting the Site Survey
As you complete every cell, stop and document the coverage
area, AP location, and antenna type. You can use this information for notes for
the final documentation (discussed in Chapter 13); this information also proves useful as
you move to the next cell area to survey, helping you judge well where to place
the AP to start the survey process for the next area. This is also a good time
to make any notes about issues or possible problems that you discovered but were
not included in the pre-site survey documentation.
Sections
