Conducting a Wireless Survey
Recently, my employer was approached to perform a wireless site survey for an international company with a distribution center in the area. Because I am the person with the most wireless knowledge in the company, I was asked to lead this project. This is a summary of the steps we took in order to assist our customer with a successful wireless implementation.
When we first met with the company, it explained the situation: It was looking to implement a new voice-activated system that pickers would use for receiving notification of what needed to be added to each order. The purpose of this new system was to provide increased productivity in the warehouse operations. In an effort to put this system into service, members of the network team from the corporate office performed tests and deployed what they thought would provide the required coverage. Sadly, their testing proved to be futile, thus instigating the request for us to perform a full wireless survey.
Our initial meetings brought to our attention many obstacles to a successful wireless implementation. First, the warehouse area is 100,000 square feet with about 30-foot-high ceilings, and the client wanted full coverage throughout the warehouse at all levels. Further, because of the nature of the warehouse and the products stocked there, many obstructions exist that would greatly affect the coverage areas. Finally, the devices the client wished to use provide limited range because of to power limitations.
With these requirements in mind, I set about to prepare the materials necessary for the survey. Even though I had performed other surveys, this was the first of this magnitude. Thus, I wanted to be sure I had the proper equipment. The following are the items I included in my wireless survey kit:
- A small cart to facilitate movement of equipment
- Two wireless access points for testing purposes
- A variety of antennas (diversity patch, column mount, Yagi, ceiling mount and omnidirectional)
- Duct tape
- 3M Command adhesive strips
- One long, grounded extension cord
- A small UPS for providing power in areas away from power outlets
- Laptop and PDA with wireless card for testing signal strength (In this case, we used the discontinued Cisco Aironet 350 PCMCIA card because the devices only support 802.11b)
- Sheet protectors
- Dry-erase and permanent markers in various colors
- Network cables to connect the access point to the power injector
- A scissor jack to simplify mounting the access points at the proper height
- Multiple copies of the map of the survey area
Armed with these items, one co-worker and I went to work surveying the area. We started by configuring the access points and wireless devices. In order to best duplicate the devices the customer would be using, we set the power on the wireless cards to 30 mW, per instructions from the vendor. The power level on the access points also was reduced to 50 mW to diminish the chance of interference by two access points on the same channel overlapping. This is due to the fact 802.11b has only three nonoverlapping channels — 1, 6 and 11. In addition, while the devices might be able to receive a signal further out, they would not be able to respond at that distance, so the additional power on the access point would provide no useful purpose.
Further, we determined what level of service we should set as the minimum required. The wireless device’s documentation recommended limiting the speeds to 5.5 mbps and 2 mbps, but we chose to limit our connections to 11 mbps and 5.5 mbps instead. We also set a minimum signal-to-noise ratio of 25 dB and a maximum packet loss of 3 percent to ensure a good-quality connection.
Following a quick walk-through, we decided on a starting location and antenna, mounted the antenna, then powered up the access point. We continued to the next location and performed the same steps. This would provide us with two access points to test, thus helping to reduce the time necessary for the survey by allowing us to test coverage for both access points at the same time.
With the two access points in place, we started the actual survey. Because the customer had requested full coverage at all heights, we tested signal strength at three levels. Starting at the ground level, I walked around with my PDA and tested signal strength and quality of connection to both of the access points we deployed. With the map in a sheet protector, I placed a dot on the map with a dry-erase marker where the signal dropped below the standards, connecting the dots as we went along. We then repeated these steps for each access point and level, using separate sheet protectors for each level and access point. Once we determined the location was a good choice, I redrew the coverage area with a permanent marker, using a different color for each access point, to keep from erasing what had already been tested.
After determining the final location for the first two access points, we continued these steps, placing the access points in various locations throughout the warehouse and testing the coverage from each location. This entire process took about three days because of the large area being tested. In the end, we determined 11 access points would be necessary to provide the coverage requested. Based on this number, a wireless LAN controller also would be implemented to eliminate the need to configure each access point individually and provide better control.
Now that I had the data, it was time to create the final documentation. While the sheet protectors with the coverage areas drawn with permanent markers were functional, it didn’t meet the quality standards that I wished to provide. Therefore, I used a digitizer table and created a digital version of the coverage areas. This took a great deal of time and effort to complete, but in the end, it was worth it.
A separate graphic file was created for each of the levels tested, with individual layers and colors for each access point. This would allow me to print each access point’s coverage area on a transparency, thus providing the ability to look at each access point separate from the others while still maintaining one file that showed the overlapping coverage areas provided by each access point at that level.
All this documentation was combined in a three-ring binder along with a summary of the findings and the methodology used. The binder included 8.5-inch-by-11-inch versions of the coverage maps and transparencies for each level, with 11-inch-by-17-inch versions for each level showing the combination of all the access points for that level. Each of the coverage maps included a legend, showing which color corresponded to which access point.
Finally, the binder was presented to the company, we reviewed it to ensure the customer understood what was provided. The customer chose to deploy the access points itself, so every step was taken to ensure it had the necessary information for successful implementation.
With the survey completed, the customer took over. Once it received the recommended equipment, the customer started the implementation. Unfortunately, after completing the deployment of all the access points, it was still experiencing intermittent lost connections, so I was asked to come back to help troubleshoot the problems.
On my return visit, I found most of my recommendations had been followed, with a few exceptions. The first exception was with the placement of the antennas —I found three of the antennas were rotated 90 degrees, thus reducing the coverage area provided by these access points.
The next exception was with the power levels configured on the access points. Instead of setting the access points to the power level of 50 m