How We Tested Ideally, to test any outdoor wireless product, it should be set up outdoors. Unfortunately, we weren't able to test these bridges that way. In addition to the fact that we had numerous wireless bridges with various distance limitations, it wasn't practical to set up permanent antennas on two buildings several miles apart. Instead, we tested the bridges indoors and used attenuators to simulate distance and signal loss. Our objective was to generate enough traffic to flood the wireless bridges with data and measure their performance under variable attenuation. We connected each bridge to an Intel Express 10/100 switch. Two Dell OptiPlex GXA machines, one running Microsoft Windows 2000 and the other running Windows 98 SE, were connected to the Intel switches. Variable attenuators were used to connect the bridges, and to provide attenuation for our range testing. These two network segments were connected via wireless bridges through our attenuation equipment. We used attenuators to simulate the signal loss experienced in long-range installations. To evaluate data-throughput characteristics, we installed Ganymede Software's Chariot network-testing system on both Dell machines. Using Chariot, we ran various tests, including variable-file-size FTP and Web-based traffic. We then varied the attenuation from nearly 0 dB to 40 dB and ran the tests at different attenuations. Ascertaining the level of attenuation at which each bridge ceased to function let us determine range. Telecom Analysis Systems lent us a TAS 4500 Flex5 RF Channel Emulator, which can emulate multipath fading and delay spread, and log normal shadowing. This device let us perform enhanced testing and get a better sense of how the bridges would react under different conditions.