Best Case Throughput
For each AP/STA pair, once the products were connected into the Azimuth system, I started by running the IxChariot throughput.scr script with TCP/IP and the file size changed to 1,000,000 Bytes from its default of 100,000. I then ran multiple tests to get a feel for how each product behaved.
I ran separate upstream and downstream tests and in the case of the products that had only 10/100 switches, ran tests using two wired clients. But I settled on using simultaneous up and downstream test runs with 0 dB attenuation to test the maximum throughput of each product.
The Azimuth system also has inherent path loss, even though it uses direct cable connection, but it is usually small enough that some signal attenuation must be used so that receivers under test are not overloaded. The test configuration I used had 44 dB of fixed attenuation before any programmable attenuation was applied. So the 0 dB attenuation referred to above is referenced to the fixed system attenuation.
Note: An AP and STA sitting 5 or 6 feet apart in open air will have around 40 dB of signal attenuation between them. This is referred to as Path Loss.
Figure 12: Buffalo Infiniti 0 dB Uplink/Downlink Throughput
Figures 12 and 13 show the maximum throughput of the Buffalo Infiniti and Linksys Wireless-N product pairs. Although both are based on Broadcom Intensi-fi silicon, the Buffalo products averaged only 86 Mbps, while the Linksys just barely tops the 100Mbps mark at an average total throughput for both up and downstream of 101 Mbps.