The results of our QoS and filtering tests are summarized in charts . If the product performed perfectly, it received a "perfect" mark.

PDF formated charts of How We Tested
QoS Test Results and Filtering Test Results

· Quality of Service Tests To test QoS, we used a pair of Optivision LiveSystem video encoders to create a 10-Mbps stream of real-time MPEG-2 video, which was fed across the network from a Pioneer DVL-909 DVD/Laser Disc player. Videowas chosen because of its high sensitivity to fluctuations in the quality of the network connection.

In addition, we used an Ixia Communications Model 1600 traffic generator to generate TCP/IP background traffic, saturating the 100-Mbps link between the two switches. We also checked the switches' ability to assign high priority to traffic coming in on a specific port. Next, we tested each switch's ability to assign priority to a packet based on source or destination IP address.

· Filtering Tests We tested each vendor's ability to filter based on MAC (Media Access Control) address, source and destination IP address, and Layer 4 TCP/UDP port number. Extreme supported MAC and destination IP address filtering, while Foundry supported filtering on source and destination MAC or IP address or on Layer 4 port number.

· OSPF Routing Functionality To build a fault-tolerant Layer 3 network, OSPF routing functionality is a must. We verified that each switch we tested could operate in a dynamic OSPF environment for at least 72 hours without failing. Using test equipment from Midnight Networks, we created a simulated routing environment consisting of 50 routers and more than 1,600 external routes. Each product was connected to the Midnight Networks Avalanche tester, and it was allowed to run for 72 hours uninterrupted. The Avalanche unit simulated a dynamic OSPF environment with numerous changing routes. All products we tested passed unscathed.