One of the primary responsibilities of a network operator is to monitor the network and bandwidth utilization. This may be for various business purposes, such as:

• Ensuring that low-priority or scavenger traffic is not choking the network and making business-critical traffic suffer
• Ensuring that the WAN provider is able to provide the committed traffic rate on the access circuit
• Use the result for capacity planning to upgrade or downgrade the bandwidth

In most of the troubleshooting and analysis recipes we have discussed so far, the data capture will be done as close as possible to the application endpoints. In order to measure WAN circuit utilization, we may need additional consideration to measure the bandwidth accurately. In the preceding topology, WAN-Router1 may be connected through a 1 Gbps interface to a WAN provider, but the Committed Access Rate (CAR) can be much lower (such as 10 Mbps). Normally, WAN providers will have traffic rate limiting enabled in the inbound direction to limit the traffic to 10 Mbps, and drop any excess traffic. So, performing packet capture on the WAN-Router1 side may not be accurate. On the other hand, the customer may implement features such as traffic shaping, which allows them to buffer the traffic and ensure that the outbound traffic from WAN router to provider does not exceed the CAR. So, depending on how the network is deployed, packet capture will need to be done at the appropriate position.

In this recipe, we will discuss how to use Wireshark to measure the WAN circuit bandwidth and analyze different application traffic utilizations.