Use OSI Model to Troubleshoot Networks.. cool

Some of you might be thinking “theoretical models don’t help me” or “the OSI model is just some engineering thing for the nerds”. However, in reality, it is quite the opposite. The OSI model can help you.

First off, I want you to have a visual image of the OSI model. On the left hand side is a user. On the right hand side, you could have a server. Every request AND response has to travel from the left, down every layer, to the physical layer, across the physical layer, up the layers on the right, and up to the server on the top right hand corner.

By understanding this information and having a visual image of how the OSI model works, you have a very valuable troubleshooting tool. If you can visualize the different pieces that complete the path of the data, you will have much less trouble resolving a networking issue. You can think of the OSI model as a map for the path that your data takes. If your data is not completing the path, you can check each “waypoint” on the map until you find the area that is causing the problem.

Some experiences with the OSI model

In my time spent as a network administrator, I would use the OSI model daily. Let me explain how.

When a get a call from a user that explains a problem they are having, I would immediately visualize the OSI model. The user might say that they can’t bring up a graphic they are trying to download from the Internet. That graphic is brought up in an application.  (layer 7 of the OSI model). I could either start at the top or the bottom, depending on what I suspected was the problem. I would usually start at the bottom (termed the “bottom up approach”). At the bottom of the OSI model is the physical layer (layer 1). So, I would proceed to ask them questions like this:

• Is your network cable plugged in? (physical)
• Is there a link light on the Ethernet switch and Ethernet NIC? (data-link)
• Do you have an IP address? (network)
• Can you ping your default gateway? (network, testing LAN IP connectivity)
• Do you have DNS server information?
• Can you ping your DNS server? (network, testing IP connectivity)
• Do you have a firewall configured? (network on up to application)
• Can you ping the host you are trying to get to by name? (application, DNS and network WAN IP connectivity)
• What format is the graphic in? Do you have a viewer for that format? (presentation)
• Can your web browser open up another website? (basic application troubleshooting)

It may turn out that the graphic they were trying to bring up was a .TIFF file and they didn’t have a decoder for that type of file. Thus, this would have been a presentation error issue as the presentation layer deals with formats of graphics & files, as well as compression and encryption.

Methods of using the OSI model

I just gave you an example for using the OSI model with a “bottom up” approach to troubleshooting. There are three different ways to use the OSI model:

Bottom up – troubleshooting by going from the physical layer (layer 1) up to the application layer (layer 7)

Top down – troubleshooting by going from the application layer (layer 7) down to the physical layer (layer 1)

Divide and Conquer – in this method, you start with whatever layer you feel is most likely the cause of the problem, then move in whatever direction you feel is the more likely cause of the issue (either up or down the OSI model).