Customer Problem:

"We have an existing wireless network at our plant in Dallas and are experiencing connectivity issues. We are using the Motorola MC9190-G Mobile Computer for inventory control and are looking to have a site survey to analyze the coverage on site. The MC9190-G is having problems running the Web Application. We intermittently receive a web page not found error and are not able to enter data into the inventory control system. HELP!"

Wireless LAN (WLAN) Interference

In practice, interference in WLAN applications usually manifests itself as reduced data-traffic throughput, less effective range, and impaired quality of service (QoS) for voice and video applications, but can also include the complete failure of a given link or wireless network. The cumulative effects of interference may be identifiable by analyzing network management logs, but diagnosing these symptoms in this manner can be very difficult because they can also result from other network related problems. This situation further motivates the use of specialized tools for identifying and evaluating the sources and effects of interference.

With respect to WLANs, interference can come from a variety of sources. Interference from other WLAN networks is typically co-channel interference (CCI), usually between two access points on the same channel, or adjacent-channel interference (ACI) resulting from two access points operating on abutting or overlapping channels. Since WLANs employ a “listen-before talk” protocol, based on Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA), any interference between WLAN networks tends to work out somewhat cooperatively, with the two networks often sharing channel capacity as noted above. In contrast, interference from non-WLAN sources, which use protocols different from those of WLANs, more often result in the degradation of WLAN transmissions. There are numerous non-WLAN devices that operate in the unlicensed bands, including Bluetooth products of many forms (some operating at the same power levels as WLANs), cordless phones, wireless video surveillance cameras, wireless security and energy management systems, proprietary wireless bridges, and computer peripherals such as cordless mice, keyboards, and game controllers. In addition, there are emissions from commercial and industrial devices such as microwave ovens, certain RADAR systems, and even microwave-based lights.

One approach to dealing with interference is to move to another band, most obviously the 5 GHz spectrum used for 802.11a and 802.11a/n. RHO Wireless recommends deployment using both the 2.4 GHz band of 802.11b/g/n and the 5 GHz band of 802.11a/n and the next generation 802.11ac. Not just because the 5 GHz spectrum is currently less likely to suffer from interference; but, there are 21 non-overlapping channels defined in the 5 GHz WiFi spectrum in the US (as compared to just three non-overlapping channels in the  2.4 GHz band). This offers significantly more uncongested capacity. The 5 GHz spectrum has been underutilized primarily due to a lack of familiarity, and a general belief that transmissions at 5 GHz have less range than those at 2.4 GHz. While it is true that 5 GHz signals do not propagate as far as signals at lower frequencies, the throughput increase of using of 802.11a/n networks brought about by channels bonding and multiple input and output streams far exceeds that of 802.11g and 802.11g/n networks.

Problem Analysis

The first thing we asked for was a site plan showing the deployed access point locations, the type and model of the access points installed, the access point operating channels and the access point operating power levels. The installation was completed 6 months previously and the wireless network had "never met the our expectations". The wireless installer had not completed either a pre-deployment nor post deployment site survey.  The installer "just placed the access points where they thought they were needed". Unfortunately, this mistake is made far too often and may result in coverage holes, or, more likely, severe co-channel interference. The customer was using Cisco AIR-AP1141N-A-K9 which is a single-band Standalone 802.11b/g/n Access point which operates in the 2.4 GHz band. Thus, there are only three non-overlapping channels available for use in the wireless network.

The customer's plant was using half of a building with the other half reserved for plant growth. Total operating space was 267,000 square feet with wireless coverage provided by eleven access points. Since there are only three non-overlapping channels and there are eleven access point, there was a strong potential for co-channel interference from eight of the eleven access points; plus, the the interference introduced by the customer's neighbors' WiFi networks.

We conducted a Predictive Analysis of the customer's facility using the AP locations, power settings and channels provided by the customer. The wireless installer had not provided a map of the AP locations and the customer could not locate one access point. So we "guessed" at the location of the missing AP on the the virtual site survey used for the Predictive Analysis.

Predictive Analysis Results

Predictive Analysis signal levels showing excellent signal strength and coverage

Predictive Analysis indicated that severe Co-Channel Interference was to be expected

But, what is the real coverage? What's the real-life performance? Are there coverage holes? Are there really interference issues?  What's the interference from the surrounding area? Only a Post Deployment site survey can provide the answers.

Post Deployment Site Survey

A post deployment site survey was completed with results that mirrored the Predictive Analysis except:

• Two additional customer 2.4 GHz band access points were discovered. They had been installed by the customer prior to the implementation of the wireless network plant wide.
  • Since no previous site survey had been conducted, the legacy access points had not been identified.
• Actual access point locations of some of the access points was slightly different than that provided.
• Actual location of the "missing" access point was determined.

Site Survey Signal Strength and Access Point Locations

The site survey indicated that there was severe Co-Channel Interference

2.4 GHz Interference from Neighboring Facilities

Recommended Solution

Since the access points were already installed, cost precluded moving the needed access points to an optimum location or deploying a dual-band network. So, another Predictive Analysis was performed using the actual access point locations, disabling access points that were not needed, and adjusting the access power levels to minimize the predicted co-channel interference. Five out of the eleven access points were used with the new power settings implemented. Eight access points were disconnected from the network (Six Cisco AIR-AP1141N-A-K9 access points and two legacy access points).

Recommended solution predicted signal levels showing signal strength and coverage
that meet the customer's requirement

Recommended solution showing the Co-Channel Interference
was minimized to an acceptable level

Conclusion

Wireless network installation is a complex task that involves many invisible and unpredictable variables such as RF signal strength, noise level, data rate, channel allocation, user density, and so on. Since all these factors ultimately affect wireless LAN network performance, you should always have a site survey performed before installing your wireless network to fully understand the behavior of radio waves within your facility. This will minimize your installation costs and ensure that you have a wireless network that meets you expectations.

RHO Wireless uses the Fluke AirMagnet Site Survey Pro software application and our experience to determine the optimal number and placement of wireless LAN access points.

- - Have AP, Will Travel! - -