Recommended Facility Documentation
Any engineer surveying and installing a WLAN system needs
facility documentation, the collection of which should be part of the initial
design stage. Facility documentation enables the engineers defining the WLAN, or
trying to survey and install it, to identify areas of concern, areas of
coverage, user densities, and even types of antennas to consider, before ever
walking onto the site. After the survey has been completed, the facility
documentation becomes a critical part of the overall documentation and should be
kept current as to any changes and maintained with the rest of the network
documentation for future reference. The facility documentation should include
(but is not to limited to) the following:
Site Map
Before beginning any site survey, obtain a good site map or
floor plan of the site. In some cases, the site map might not really be a floor
plan, but more of a site layout, including building layout and contents as well
any outdoor areas that will be covered.
This site layout document will become part of the final survey
and network documentation. As such, having a soft copy of this document is very
helpful so that it can be copied and distributed as necessary to the survey
engineers, installation team, and network support staff. Extra copies of the
site map should be available to make notes on during the actual survey and
installation steps, as well as during any presurvey discussions.
Prior to the survey, you can use the site map document to
define desired coverage areas; identify where coverage is not needed; and define
user locations and densities, problems areas, network closets, cable runs, and
plenum areas. Basically, this document becomes the physical schematic of the
wireless network (see Figure 8-1).
Building Construction
Building construction can vary widely from site to site.
Materials and construction techniques in San Francisco differ significantly from
those used in New York City, London, or Cairo. Differences in building
construction, even though sites might look similar, can cause RF to react in completely
different ways. Figure 8-2 shows examples
of various building materials.
A multifloor building might use precast, reinforced concrete
for flooring. Although this type of construction might create some attenuation
problems for RF, the effect on RF penetration is significantly less than it
would be with a floor of poured concrete over a steel pan. Although some RF may
get through in this latter example, the steel pan provides a very good RF shield
between floors (see Figure 8-3).
Walls can be similarly deceiving. In most industrial buildings,
it is common to use steel studs with drywall or plasterboard over them. The
drywall and plasterboard cause only a slight attenuation of RF signals, and the
placement of the steel studs has little effect at all. Other walls might be
concrete block, with or without steel reinforcement, which cause only limited
attenuation of RF. However, precast concrete, typically using steel
reinforcement, is a different story. The amount of steel used for reinforcement
inside the concrete will cause the RF attenuation to vary from one building to
another.
Although drywall and plaster usually minimally affect RF, the
material behind the wall can pose problems. Consider a real example from a
health-care facility. The RF energy was having a hard time getting into several
offices. Further questioning of maintenance personnel at the facility and
reviewing some older building documents revealed that this area had been
remodeled recently. Before the area was used as offices, it was the radiology department.
The x-ray room had been turned into offices. And, as typical with x-ray rooms,
the walls were shielded to prevent x-ray energy from leaking out of the room.
The walls were not removed, just covered over; therefore, the RF could not get
into the offices.
In some buildings, the walls might be made from a form of
reinforced wire mesh, with a plaster-type material spread across it (often
called stucco). The mesh can work much like an RF
screen, causing a severe level of signal loss or RF attenuation.
Steel outside walls, or steel walls separating parts of a
building, can detrimentally affect RF coverage (because the wall might not just
restrict RF penetration, it might also create a large number of multipath
signals). This is common in industrial facilities, where a building has
undergone one or more additions. What was once the outside wall might now be a
partition between the old and new sections of the building, causing both
multipath signals and an RF shield between building sections.
Be sure to research this information before or during the
survey and document your findings on the site map.
Building Contents
One often-overlooked area of concern is the building contents.
Those with minimal WLAN and RF experience sometimes underestimate the effect
that building contents can have on a WLAN.
Figure 8-4 shows several
examples of problems that can occur in a typical office environment. Areas such
as file rooms and storage rooms are often filled with steel cabinets, creating a
very large RF shield for RF entering that room, or even passing through it to
other areas of the facility. Although most would assume an area filled with
cubicles should have minimal effect on RF, it might in fact create a challenge
for RF coverage. The number of cubical partitions, the amount of steel in the
partitions and desks, and the size and make-up of the bookshelves can affect RF
range.
Another area that is very difficult to cover is a library or
documentation area. Shelves full of books are shelves full of paper, and most
paper has a high level of attenuation to WLAN frequencies. It is very common for
WLANs to use directional antennas, focusing the RF energy down the aisles of the
books. (See Figure 8-5.) Because of
similar shelving, warehouses and even some retail stores also use directional
antennas in this way.
Kitchens and break-rooms usually contain microwave ovens. Microwaves are
also found around many health-care and industrial facilities for purposes other
than heating food. Although microwaves pose no problems for 5-GHz WLANs, they
can be problematic for 2.4-GHz WLANs. The typical microwave oven uses the same
frequencies as a 2.4-GHz WLAN. (This is because 2.4 GHz is the resonant
frequency of water, and when 2.4-GHz energy strikes water molecules, it is
absorbs the energy and causes the molecules to vibrate, creating friction and
heat.) Locating a 2.4-GHz access point (AP) close
to a microwave can cause undue interference and result in poor RF
communications. Take care to keep these APs (and clients when possible) at least
10 feet away from any standard microwave oven. It is therefore recommended to
note the location of any such devices on the site map. Also be aware that
industrial microwave ovens sometimes have a much higher power than those found
in the home or office, possibly creating even more interference. Testing should
include RF coverage verification while any microwaves in the local vicinity are
in full operation.
In one case, a health-care facility was having trouble with
one particular AP that was dropping all associations intermittently. Close
inspection of the facility turned up a microwave oven in an area that was not
part of the RF-covered area, but was located in a lab adjacent to the AP-covered
area. The problem was that the oven was located on the other side of the wall
(made from drywall) from the AP, with a total of about 5 feet (and two pieces of
drywall) separation. This is why it is important to understand the entire site,
including areas where coverage may not be needed. Figure 8-6 shows a site map with potential problem areas
labeled.
Locations such as emergency rooms and cardiac care in hospitals
use sensitive equipment such as electrocardiographs (EKGs) and other monitoring
systems. Although these devices are not generally a problem with WLANs, take
care to locate the radio gear near the devices and to verify that the RF in the
area does not cause any interference. One common problem occurs with older
plotters and printers. The RF energy can cause slight variations in the print
and plotter driver mechanisms, resulting in "glitches" in the patient
printouts.
As discussed in Chapter 6, "Preparing for a Site Survey," all cordless
and wireless devices should be inventoried. Phones, speaker, cameras, cordless
mice, cordless keyboards, baby monitors, and virtually anything that might be RF
related should be noted.
In a warehouse, retail environment, or even an office building,
a change of contents can greatly affect the coverage of an AP. Inventory levels
often change in a warehouse or retail facility. At certain times of the year
(such as early November, when stock levels rise for the holiday shopping
season), stock levels in some facilities may reach beyond 100 percent, with
material placed in any possible free space, such as directly in front of the AP
that provides coverage to the area. This poses a real problem for the survey
engineer who is trying to survey when the stock level might be at a low level
corresponding to the season. (Many installations occur during the off-season,
when facilities are not running at peak capacity.)
Defined User Areas and Densities
The topic of user density has been brought up many times in
this book. As stressed previously, defining user areas and densities is a
crucial part of the design and must be on the minds of design engineers and
survey engineers at all times. The overall performance of the WLAN system
depends on proper user density.
There have been surveys based on nothing but user density. At
one very large software company, the buildings were all built in a very similar
manner, and with identical internal design and contents. All cubicles were
identical, all office construction was identical, and the number of users in a
given area was very similar.
For this customer, it was decided that the applications used by
nonengineering employees would permit between 20 and 25 users per AP. This
provided adequate performance for normal operational network load. The
engineers, however, required a bit more performance, and the user density was
lowered to between 10 and 15 users per AP.
Based on information such as this, some of the design can be
done up front. You can use the site map to determine how large the cell coverage
needs to be. For example, a survey determined that a single AP set to default
power levels, with dipole antennas, could provide coverage for many more users
(based on their seating locations) than the design calls for. In this particular
case, the desired coverage turned out to be a small circle on the site map, and
was about the size of a coffee cup. From this point, it was a matter of defining
how many "coffee cups" were needed (see Figure 8-7). The engineers then selected the power setting
to provide the proper coverage for the user density in the appropriate areas.
Finally, testing was completed to prove the guesstimations of the coffee cup
survey.
If voice is to be
used over the WLAN, it is vital to the design to understand the capacity of the
AP versus the number of calls that can be carried by one AP at any given time.
Typically, a standard 802.11b AP supports only between six and eight calls with
the standard compression used in 802.11 voice over IP (VoIP)
phones. As compression techniques improve, or as the wireless 802.11 VoIP phones
move to support 802.11g or 802.11a, the number of calls per AP will
increase.
Sections
