Network Troubleshooting

When troubleshooting network problems, it’s very important to keep theOSI model in mind and work your way up from the physical layer to the application layer. This bottom-up approach helps to successfully troubleshoot network problems because each layer relies on the lower one to function properly. In this article, we’ll provide some basics on troubleshooting the first three layers of the OSI model.

Troubleshooting the Physical Layer (Layer 1)

The physical layer includes anything that generates and moves bits from point A to point B. This is the network interface layer – such as Ethernet or WiFi cards, fiber cables and the air that enables hosts or computers to communicate with other hosts and the outside world in general. To troubleshoot this layer, the network engineer can use the diagnostic tools that the vendors include in their hardware.

In the case of Ethernet cards, basic diagnostic commands report information on the duplex and link speed that the card has established with the other side of the cable. In the case of a WiFi adapter, the utility should report the signal strength and link quality of the connection established with the base station. This data is important to understand the quality of the layer 1 link established.

To troubleshoot problems with copper or fiber cables you can use a time-domain reflectometer (TDR) or optical time-domain reflectometer in the case of a fiber link. Some networking vendors also include basic TDR function on their equipment. In the case of WiFi networks, spectrum analyzers are very useful to provide information about “the air” and detect any interferences in the surroundings, such as microwave ovens. You can refer to vendors such as MetaGeek or Ekahau for such solutions.

Troubleshooting the Data-Link Layer (Layer 2)

To troubleshoot the data-link layer issues, network engineers can access the command line of a switch to inspect the MAC address table, which provides information about the MAC addresses learned on switched ports. To troubleshoot Layer 2 communications between hosts, network engineers can use passive analysis tools such as wireshark, which is GUI based, or tcpdump, which is command line based. Such tools provide a recording of frames, flowing across a network link, switch or host.

Another important thing to keep in mind when troubleshooting layer 2 issues is spanning tree. Spanning tree is a Layer 2 protocol that enables switched networks to build a loop-free topology, which happens when redundancy is introduced in a network design. When a network topology has a loop, frames flow indefinitely without reaching its destination host or getting discarded, causing broadcast storms. Broadcast storms saturate network links and cause instability in the CAM (Content Addressable Table) of switches. The spanning tree protocol avoids this scenario by disabling switch ports that cause loops. However, for spanning tree to properly work, all switches in the network must be correctly configured. Getting familiar with spanning tree and diagnostic commands on switches is a very important knowledge for network troubleshooting.

Troubleshooting the Network Layer (Layer 3)

The most used commands to troubleshoot layer 3 issues are ping and traceroute. With ping you can verify whether a host can reach a destination network or host. With traceroute you can discover the routing hops available between a source and a destination. When troubleshooting layer 3 problems, it’s important to consider whether the destination host is located within your organization, or not. If it does, then the troubleshooting efforts aim at figuring out whether a network misconfiguration, or something else, is causing the connectivity or performance issues. If, on the other end, the network path to the destination host traverses a third party, then it’s important to provide enough information and prove that it’s someone else’s problem. One way or another, ping and traceroute are two useful commands that shed light on reachability and performance issues between two IP hosts.

Closing remarks

When troubleshooting network performance issues, it’s very important to keep in mind the OSI model and its layers. Starting from the bottom layers and moving your way up will assure that the proper troubleshooting procedures with faster problem resolution.