What metrics and tools can accurately measure network performance?

Tracking network connection performance is fundamental to identifying and eliminating potential bottlenecks. IT administrators can examine several performance metrics to measure network performance and get a clear picture of service quality over time.

The following performance metrics can help shed light on issues that delay or interrupt traffic flows:

• bandwidth
• throughput
• jitter
• latency
• error rates

First, IT professionals need to understand how much bandwidth is available and how much traffic their network should theoretically carry in a given time frame. A Gigabit Ethernet connection, for example, can carry 1,000 Mbps or 125 MBps.

Several factors can depress the throughput or actual rate at which data moves through the network channel. This makes it important to measure throughput on a consistent basis to spot potential issues early. 

One way to measure throughput is to send a file from one endpoint to another. However, this may not be the most accurate way to measure network performance, because -- along with tracking how fast the file crosses the network -- this measurement also includes hard-disk, operating-system and other delays that happen at the source and destination endpoints.

Whatever methods an IT group uses to measure network performance, the methods should be applied on a consistent basis.

Depending on their resources, IT groups can use high-performance servers or workstations that solely measure throughput during the testing process to avoid these bottlenecks. Other tools and methodologies can also extract the most accurate throughput measurement. Tools such as IxChariot, iPerf, Spirent TestCenter and NetCPS can test and report transfer times.

Some of these tools can also measure network performance by evaluating other performance metrics, such as latency and the time interval between when an instruction is issued to transfer data and when that transfer is actually initiated.