To test interoperability, we installed all access points and PC Cards, initially using the default manufacturer's settings. (The ESSID was the exception--it had to be set to the same value for all products.) We then configured Ganymede Software's Chariot network performance management software to transfer 250,000-byte files using FTP for one minute, from a Compaq Armada 7800 notebook running Windows95 to a Micron Milennia Pro2 Pentium Pro 200 server running Windows NT 4.0. All tests were performed on an isolated Ethernet segment. Table 1 shows the results for direct-sequence products. Table 2 illustrates the performance of frequency-hopping products.

PDF CHART
DS Interoperability: Single-Station FTP Performance (Mbps)

With the exception of MaxTech Corp.'s PC Card, the products were interoperable, though in several cases, we had to consult with vendors, alter default settings and upgrade firmware and drivers. Note that our interoperability tests with Intermec's AP were performed using a beta unit. A replacement unit shipped to us just prior to our deadline turned in performance numbers comparable to other units we tested.

PDF CHART
FH Interoperability: Single-Station FTP Performance (Mbps)

We measured multistation performance to determine how wireless devices operate under contention and the degree to which stations close to an AP perform better than stations further away. To test this, we installed PC Cards in two Compaq Armada 7800 notebook computers and two Dell Latitude CP notebook computers. One of the Armada notebooks and one of the Latitude notebooks were located in the lab within 20 feet of the AP. The other pair of notebooks were in an office approximately 75 feet away from the AP. Using Chariot, we transferred 250,000-byte files via FTP for five minutes concurrently between each notebook and our NT server. We then measured the average FTP performance for each notebook, the aggregate throughput of all four stations and the average station throughput for each product.

PDF CHART
FH Multiple-Station: Near and Far Nodes, FTP Performance (Mbps)

PDF CHART
DS Multiple-Station: Near and Far Nodes, FTP Performance (Mbps)

To evaluate each product's transmission range, we established three test environments in the Center for Science and Technology at Syracuse University, a 10-year-old office and classroom building. We set up APs in a hallway and ran single-station FTP tests between two notebook computer systems. Although the FTP client we used reported performance slightly higher than Ganymede's Chariot software, we were more concerned with each product's relative performance.

Location A, used to simulate a long-distance, open-office test environment, was approximately 250 feet from the AP down a long corridor. Location B, resembling a short-distance, walled-office test environment, was in an office approximately 100 feet from the AP, but separated by numerous walls and corridors. Location C, illustrating a long-distance, walled-office environment, was in a classroom approximately 150 feet away from the AP, again separated from the AP by numerous walls and corridors. Table 5 shows the results for each of the products. The performance numbers are those reported by Ipswitch's WS-FTP client.

PDF CHART
Our test Environments: Testing Transmission Range (Mbps)