Finding Waves
##AUTHORSPLIT##<--->
Techniques for a Successful Wireless Site Survey
Wireless local area networks are the most widely adopted networking technology to hit the market in the last three years. They have the potential to make network applications and the Internet available anywhere on a campus so that students and faculty are no longer tethered to their offices or shared computer laboratories in order to connect to a computer network.
In early 2001, Saint Francis University in Loretto, Penn., undertook a study to create a wireless network in all of its academic buildings. The Center of Excellence for Remote and Medically Under Served Areas (online at www.cermusa.org), a telemedicine and distance learning research facility located on the university's campus, collaborated with Saint Francis on the effort. The university's goal was to create a wireless network to support an upcoming laptop initiative in which all incoming freshmen would receive notebook computers. CERMUSA's objective was to gain a better understanding of wireless technology, as well as to create a best-practices recommendation for schools and universities in the techniques of deploying a campus WLAN.
After a year of researching different methods of WLAN deployment, three different procedures that can be used to install a wireless network were identified. CERMUSA termed these three methods: E-Witching, Scientific Divining and Computer Modeling.
E-Witching
Electron witching, or e-witching, is akin to the approach used by early settlers in the Midwestern United States to find water sources for homes and farms. On these arid landscapes, visible sources of water such as streams and brooks were hard to find; however, potable water sources were often discovered flowing beneath the surface of the ground and a well had to be dug to capture the water. Skilled settlers, known as diviners, were often called in to locate these hidden water sources by using a forked stick (i.e., a divining rod) that would bend down when moved over a possible water source. The driller would then sink a well at this spot, most often with positive results.
E-witching works in a very similar manner. CERMUSA technicians used the software included with a wireless network card to measure the signal strength to various locations on campus. When it was determined that signal strength was too low to maintain a strong connection, a new access point was placed in the immediate area. Once a building had enough access points to provide sufficient coverage for the entire structure, the network administrator would use his software to individually configure each access point using a commonsense approach in regard to channel allocation.
Measuring signal propagation using the e-witching technique is a simple procedure. To begin, the site surveyor needs the following three pieces of equipment: a wireless access point; a wireless PC card installed on a laptop; and signal measurement software, which is typically included with a wireless card. After the wireless card and software have been installed on a laptop, the first task is to locate a starting point within the area where the WLAN is to be installed. This spot should be somewhere near the center of the proposed coverage area, and should be accessible for network and electrical wiring to be run to an access point.
Once the access point is in place, simply plug the device into an electrical outlet; a data connection is not necessary at this step. Then, use the signal measurement software included with your wireless card to begin measuring the signal propagation of the access point. The software will typically display a bar or line graph to indicate statistics such as signal strength, interfering signals ("noise") and signal-to-noise ratio (SNR). In addition, most of the software quantifies signal strength with general ratings such as "Good," "Fair" and "Poor." For educational use, signal strength of at least "Fair" is needed to ensure a sufficient signal.
To begin e-witching, use the laptop with the signal measuring software to map the perimeter's usable signal. Begin by walking away from the access point while viewing the signal strength on the laptop. When the signal degrades to a point where it is fluctuating between "Poor" and "Fair," mark that location as being the perimeter of signal propagation from the first access point. Remember, the access point's signal d'es not follow a flat plane, and the signal from the access point can usually be measured on the floors both above and below where the access point is placed.
After the perimeter for the first access point's signal has been marked, examine the perimeter line and determine if coverage is adequate for the space that needs wireless coverage. If not, find the most sensible location to place the next access point that is beyond the perimeter line and move the access point to this location. Perform the same site survey with this new location and mark the perimeter of this access point location. Continue following these steps until enough access point locations have been surveyed to provide complete coverage.
Once the site survey is complete, the surveyor knows the exact number of necessary access points as well as their locations. The final step is to purchase the access points for installation and configuration into the existing hard-wired network. E-witching is the simplest technique for performing a site survey. Depending on the size of the facility, this technique can typically be completed in just a few hours; however, very little documentation is delivered in this method. Another drawback with e-witching is that there is a tendency to "overnetwork" or place an inflated number of access points in the facility.
Scientific Divining
Performing a site survey using the scientific divining method overcomes the pitfalls of e-witching. Documentation is generated through the use of coverage maps specifying signal strength in any location. A complete scientific divining survey will also help eliminate redundant access points using a scientific method to determine the best locations for access points. The scientific divining method requires identical equipment as the e-witching method, with the addition of a fire-escape map, which is basically a quick drawing of the layout of floors that are to be covered by the WLAN. If these diagrams are not readily available, they can be drawn by hand.
The first step in scientific divining is to subdivide the map into measurement zones (each zone is a logical room breakdown). For example, a small office may be designated as Zone 1A, while a larger room may be split into Zones 1B and 1C. For each floor of a building, a spreadsheet should be created that lists all of the zones in the top row and allows for written entries of signal strength in the row immediately beneath. Once the map has been subdivided into specific zones and the spreadsheet created, it is time to begin the actual site survey.
The steps involved in the site survey using scientific divining are identical to the steps used in e-witching with one exception: The surveyors are now responsible for writing the signal strength for a zone on the spreadsheet. Several access point locations should be surveyed when using this method. If time allows, every location that is a candidate for placement of an access point should be evaluated. For each access point location, the surveyor must use a new printout of the spreadsheet and note the location of the access point during the survey.
Once all access point locations have been surveyed, it is time to determine the most efficient placement. Use the data collected on the spreadsheet in combination with the map(s) to determine which combination of locations provides the best coverage for your needs. The goal is to cover the area with the fewest number of access points. With the data from every survey, a composite of the most suitable locations can be created to visually show which areas will be covered by which access point.
The advantage of scientific divining over e-witching is that documentation is created which shows coverage is complete. In addition, it may be determined that fewer access points were needed than originally thought. The disadvantage of this method is that it requires notably more manpower and time to complete than e-witching.
Computer Modeling
The e-witching and scientific divining methods of performing a site survey are ideal for small, contained environments that have the necessary resources to perform complete surveys. However, when dealing with larger campuses with several buildings, computer modeling may be the most efficient method of performing a site survey. Computer modeling involves creating virtual representations of buildings within a software package and using this model to determine the best locations for the access points. The software package chosen by CERMUSA for computer modeling was SitePlanner from Wireless Valley Inc. (www.wirelessvalley.com).
The first step in computer modeling is to draw the building(s) within the software. SitePlanner allows existing CAD drawings to be imported into the software. If CAD drawings do not exist, then it is necessary to draw the buildings using the drawing tools included in the software. When a building is imported or drawn in the software, each type of building material is given a different signal-attenuation value because some building materials obstruct radio signals more than others.
Once all the floors of a building have been modeled in the software, it is time to begin virtually placing access points. The user tells the software where he or she thinks would be the best places for access points. Once all access point locations are entered, the software will run a prediction and create predicated coverage maps. The coverage maps show the predicted signal strength throughout the building using various colors to represent different signal strengths. If there are spots that need coverage where the software shows there is little or no coverage, then it is necessary to rethink the access point placement and run the prediction again.
Channel Allocation
The final step in any site survey is to configure the access points to broadcast on nonoverlapping channels. Every access point allows configuration of a channel to broadcast on. If your campus is using the 802.11b wireless networking standard, then there are 11 different channels to choose from. However, it is vital to understand that some of these channels overlap with each other. For example, Channel 1 and Channel 2 are very close to each other in the radio frequency band; therefore, wireless network cards have a difficult time differentiating between the two channels, which results in a slower wireless network.
It is possible to overcome this channel-overlapping problem by using proper channel allocation. In the 802.11b standard, Channels 1, 6 and 11 are referred to as "clear" channels. These channels have very few interfering frequencies and therefore can exist together without network degradation. When configuring a WLAN, be sure to place access points running on identical channels as far away from each other as possible.
CERMUSA found that using the three channels was sufficient for wiring its academic buildings. However, extending the WLAN into the "green spaces" between buildings without overlapping channels was not possible. CERMUSA is currently researching additional wireless standards such as 802.11a to overcome the challenge of these green areas.
Which Method Is Best?
Each of the methods has its strengths and weaknesses. The one that will work best for your facility depends on your budget, time and documentation needs. If you are simply trying to create wireless access in a single classroom or laboratory, then the e-witching method is best. However, if you are trying to create a wireless campus consisting of several buildings and you require extensive documentation, then computer modeling is the way to go. For site surveys that fall somewhere in between, the scientific divining method will provide a sufficient balance between documentation and cost to be effective.
This article originally appeared in the 03/01/2004 issue of THE Journal.