How we did it

We tested three 5520s individually and in a stack configuration. All 5520s were hardware Version 02, firmware Version 4.1.0.9 and software Version 4.1.1.02.

All tests were executed using three Spirent Smartbits 6000C chassis with 36 Terametrics XD four-port Gigabit Ethernet cards (144 ports total). All tests were run using Smartflow v4.60.003 software.

We ran three categories of tests: stack functionality, performance and management.

To continue reading, register here and become an Insider. You'll get free access to premium content from CIO, Computerworld, CSO, InfoWorld, and Network World. See more Insider content or sign in.

We tested three 5520s individually and in a stack configuration. All 5520s were hardware Version 02, firmware Version 4.1.0.9 and software Version 4.1.1.02.

All tests were executed using three Spirent Smartbits 6000C chassis with 36 Terametrics XD four-port Gigabit Ethernet cards (144 ports total). All tests were run using Smartflow v4.60.003 software.

We ran three categories of tests: stack functionality, performance and management.

In our stack functionality tests, we verified that the console port is the same for all units in a switch as a measure for assessing how well the product replicated the port information to the multiple units. To assess reboot time, we rebooted the stack from the front user interface button and measured time for the console user interface to become active. We also rebooted one switch in the stack and measure the time for the console user interface to become active. To test the software's ability to renumber the switches in the stack, we changed the numbers, rebooted the stack and checked the stack order. For failover, we verified that the base unit fails over to another switch in the stack when the base unit fails.

In our performance tests, we tested maximum throughput before loss in a full mesh traffic pattern over all 144 ports of the stack with all ports in full duplex 1000Base-T mode. This full mesh pattern was used to test every port to every other port. The throughput was done starting at 30% going to 100% at the increment level of 5% with 0.5% error tolerance. We repeated this sequence for one switch in a standalone configuration (not in a stack). This test was repeated for Layer 2 and Layer 3 traffic. We ran these tests using 64-, 128-, 256-, 512-, 1,024-, 1,280- and 1,518-byte packets.

We measured packet latency for both a single switch and in a three-switch stack. We measured latency within the stack in a full mesh traffic pattern over 144 ports at 30% of maximum line rate and up to where loss was seen. This test was repeated for Layer 2 and Layer 3 traffic. We ran these tests using 64-, 128-, 256-, 512-, 1,024-, 1,280- and 1,518-byte packets.

To test the switches management features we took into consideration how manageable the switch was from the console-based GUI, the command-line interface and the Web user interface.

Back to Clear Choice Test: Nortel BayStack 5520

Read more about lans & wans in Network World's LANs & WANs section.