Study Finds Wide Variation In Real-World Performance Of Top Enterprise Wireless LANs
The objective of the testing, commissioned by Ruckus Wireless, was to determine how current WLAN systems perform at varying locations in a real-world environment with physical obstructions and interference as well as mixed voice and data traffic.
"The wireless testing was an eye-opener," said JPA principal Al Petrick, vice chairman of the IEEE 802.11 working group and co-author of the definitive technical reference, IEEE 802.11 Handbook, A Designer's Companion. "While the systems performed similarly in the simple locations, their performance in less-than-ideal environments sometimes varied by more than a factor of ten."
Petrick noted that the industry has made great strides in bolstering WLAN security and management but has fallen behind on the fundamental issue of reliable wireless RF performance. "With the plethora of Wi-Fi devices coming on the market, more organizations will be deploying WLANs in less-than-optimal environments," said Petrick. "Delivering reliable connectivity and consistent performance to wireless users anywhere will become increasingly challenging as the air wave gets crowded and networks get larger and more complex, especially with 802.11n becoming mainstream."
Test Setup
Testing was conducted in an 18,000 square foot office space in Fremont, California using industry-standard wireless test tools including Ixia's IxChariot and Iperf. Three separate tests were conducted on the most current hardware and software from each vendor:
- TCP downlink performance on a single client in multiple locations, with and without interference;
- aggregate TCP downlink performance on 20 clients at varying locations, with and without interference; and
- voice quality measured on 19 VoIP-over-WiFi clients at varying locations with background data traffic.
Test Results
The JPA testing revealed that while most WLAN systems performed similarly at short distances (e.g. 20 to 30 feet), performance disparity widened dramatically as distance and obstructions between the AP (access point) and client increased and interference was introduced. At the most challenging location, the best-performing WLAN system beat its nearest competitor by more than twofold and outperformed the lowest-performing system by more than 15 times.
The test found that the highest aggregate TCP throughput from a single AP to 20 802.11g laptops dispersed over an average distance of 150 feet was over 15 mbps.
The JPA test also measured Voice over Wi-Fi performance with 19 simultaneous voice clients and one data client. All but one system under test exhibited fluctuating voice qualities when the AP was placed at different locations relative to the client.
SOURCE: Jones Petrick and Associates (JPA)