QoS sanity check

One way for a vendor to rack up a good score in quality-of-service tests is to "nail up" a fixed amount of bandwidth for the exclusive use of high-priority traffic. This is similar to the TDM method used in older data networks, where bandwidth is reserved and thus never available - even when there's no high-priority traffic present.

As a check against this, we ran the test a second time while offering only medium- and low-priority traffic. Vendors weren't allowed to change configurations between the first and second tests. In this second test, the medium-priority traffic should have consumed any bandwidth previously used by the high-priority class.

That's generally what happened (see 2-Class QoS chart below), Even though no switch delivered exactly the rates we desired for both traffic classes, all made use of bandwidth previously allocated to high-priority traffic. Better still, the switches generally did better with both classes this time than in the previous test with link aggregation.

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.

One way for a vendor to rack up a good score in quality-of-service tests is to "nail up" a fixed amount of bandwidth for the exclusive use of high-priority traffic. This is similar to the TDM method used in older data networks, where bandwidth is reserved and thus never available - even when there's no high-priority traffic present.

As a check against this, we ran the test a second time while offering only medium- and low-priority traffic. Vendors weren't allowed to change configurations between the first and second tests. In this second test, the medium-priority traffic should have consumed any bandwidth previously used by the high-priority class.

That's generally what happened (see 2-Class QoS chart below), Even though no switch delivered exactly the rates we desired for both traffic classes, all made use of bandwidth previously allocated to high-priority traffic. Better still, the switches generally did better with both classes this time than in the previous test with link aggregation.

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