Student Sailors are Going Wireless
February 21, 2003
The Naval Postgraduate School is in the process of deploying a campus wide wireless LAN involving both 802.11a and 802.11b. Here's an up-close look at the first step of the project: performing a radio frequency site survey of the campus.
As I've said many times in many tutorials, it's crucial that you perform a radio frequency (RF) site survey in order to detect the presence of potential RF interference with your wireless LAN (WLAN) and determine the optimum installation location for access points before deploying. Without an RF site survey, especially for larger installations, you'll likely end up with spotty coverage and possible interference that degrades the performance of the system.
As an independent consultant, I've been assisting the Naval Post Graduate School (NPS) with the deployment of a campus-wide WLAN. The NPS educates many graduate-level military students each year and supports significant research in the areas of wireless networks and homeland security. The campus, located in Monterey, California, even includes the historic Hotel del Monte, which is believed to be haunted. (I've stayed at this hotel several times and can attest that unexplainable things do actually happen, but that's another story!)
The wireless network at the NPS will eventually cover all major facilities on the university grounds. The campus includes several buildings encompassing classrooms, laboratories, a library, offices, and living quarters. The school plans to deploy both 802.11a and 802.11b (with eventual upgrades to 802.11g) throughout most of these locations in order to offer the most advanced WLAN possible. The network will provide students and researchers with a WLAN that provides access to e-mail, Internet, and local applications from just about anywhere on the campus.
NPS is taking an aggressive role of providing high performance for all classrooms and offices by ensuring that these areas have as close as possible to 11Mbps over 802.11b connections. This requires a concentration of access points near classrooms, study areas, and other places where students and faculty work. By using 802.11a/b access points with this approach, 802.11a coverage will be possible from most areas of the campus.
The deployment of the WLAN at NPS evolved from
a masters degree thesis written by Lieutenant
Commander Joseph Roth, who is now the NPS Network Security Group Department
Head. Joe is overseeing the deployment of the WLAN with a careful eye on security.
Because of the prominent status of the NPS, the WLAN will likely become a test
case and model for the entire U.S. Navy.
Initial Walkthrough/ AP Location Testing
We spent an entire day canvassing the buildings to detect the presence of potential RF interference that might impact the operation of the WLAN using an AirMagnet handheld 802.11 analyzer and to identify possible access point mounting locations. We mostly found a handful of existing access points and made notes of where they were, as well as the respective signal strengths and operating channels. The NPS does research involving radio frequencies, so noticed some potential sources of 5GHz band RF interference in a couple of the buildings.The initial walkthrough provided us with first hand information regarding the availability of electrical power and facility construction so that we could plan the next step, which was to determine the most optimum location for access points. We tested potential locations by placing an access point at a point based on our intuitions, then recorded the signal-to-noise (SNR), association status, and supported data rates within the coverage area.
We mounted the access point to a telescoping pole (using duct tape) that we extended at each location -- tt's best to raise the access point antennas to a height equivalent to the actual mounting spot in order to produce valid results. All we had to do was find a power outlet within reach of a 15 foot extension cord; there were plenty throughout all of the facilities.
All-in-all, we tested a whopping 170 access point locations in four, really long days. Our easiest building for testing purposes was the library, which contains shelf after shelf of books that didn't seem to impact radio wave propagation much. The library does contain individual study rooms, however, along the perimeter of the building. Because of steel doors, these rooms were difficult to penetrate with radio waves. These types of rooms, which are likely places where students will need access, will need access points placed close or use directional antennas to force the waves inside,.
Some of the facilities had concrete walls with steel reinforcement, a tremendous barrier to radio wave propagation. Both 802.11b and 802.11a in these buildings had a tough time making it past 50 feet. In these areas, we had to specify many more access points than other facilities with wooden or plaster construction. When possible, we located access points near classrooms and conference rooms to ensure at least the primary areas had good coverage.
Overall 802.11a (i.e., 5GHz) range was very good. In fact we found that 802.11a had range similar to 802.11b, despite what you we've come to expect from 5GHz systems. That's because we tested using an 11Mbps 802.11b range boundary. At that boundary, 802.11a generally provides connectivity on the low end at 6Mbps, at least.
A unique challenge to installing and supporting a campus-wide WLAN at the NPS is the somewhat sporadic use of WLANs for research purposes throughout most of the buildings. Installers should carefully set the access point channels to minimize inter-access point interference and maximize roaming, but the periodic presence of an access point for test purposes may cause unpredictable interference. NPS will need to control the testing of WLANs by possibly assigning a specific channel for test purposes. The problem, however, is that there are not many 802.11b channels available that allow coexistence.
This RF site survey not only provided valuable information regarding potential RF interference and placement of access points, it also gave an excellent opportunity to learn more about the propagation of radio waves throughout various types of facilities. One of my assistants for this survey was my son, Eric Geier, who recently joined my consulting firm. Additional members of the survey team were Navy personnel John Hornbrook III and Dave Watkins, both part of Joe's Network Security Group. After completing a week of testing, these guys were getting pretty good at predicting where radio waves would go. There's really no better way to learn this without extensive, first-hand experience.
Join Jim for discussions as he answers questions in the 802.11 Planet Forums.