Keeping the network operational is no mean feat for the majority of small to medium-sized businesses. Not all traffic on many corporate networks is entirely business-related. In some companies the biggest users are doing company business, but it’s usually not mission-critical.

Making sure business-critical applications get the necessary bandwidth calls for a combination of techniques and tools. Many large networking vendors feature some kind of bandwidth-shaping capabilities included in their standard product line. The problem with that is you need to know what to shape if you want to give the necessary priority to the correct things.

No task is too difficult with the correct tools. Packet capture tools have been the main tool of choice for network analysts since the installation of the first Ethernet-based networks . Network General (News - Alert) introduced the first packet sniffer in 1986, and remains the standard by which other packet capture tools are measured.

Packet sniffers sit on a network, monitor all traffic they can detect. This can be a little challenging in today’s switched network, in which all traffic does not pass over every wire. Usually, the traffic of interest is into and out of a server or across the company firewall. These points can be monitored with a variety of different techniques, such as packet capturing appliances or hardware probes.

A large amount of open-source products boast a wide range of basic capabilities for low-level examination of TCP-based traffic. One important measurement is application latency, which pinpoints end-to-end response times and specific network and server delays. This assists in the identification of specific bottlenecks that can negatively affect the responsiveness of a given application.