OSI model and Ethernet LAN protocols

I found your topic on Networking and Protocols very interesting. I understood very well the functions and operational flow of OSI model and TCP/IP on Internet.
1. Please, could you relate the OSI model to an Ethernet based LAN.
Specially, map a typical ethernet LAN to OSI Model and describe the functionalities of the LAN protocols starting from the Client to the LAN server (peer to peer or client-server)

For example, relate the protocols functions to what happens if I want to do the following on an Ethernet LAN
a). Print a document on my workstation on network printer attached to LAN server /another workstation

b). Run ms excel (network version)

I am an IT Auditor and I need more experience on protocols.
[img]images/smiles/icon_cool.gif[/img] [img]images/smiles/icon_smile.gif[/img]

Dear Sally,

It's good to hear that the Networking and Protocols section is helping you understand more on the topic. There will be a lot more information added to the site as I continue to develop it.

As far as your question regarding the OSI model and the example you asked for, they are very specific and I'll try to give you the answer you require.

I will describe to you briefly what happens in one situation because my time is very limited.

Let's take the MS-Excel networked version.

Assuming you have a windows workstation thats connected to a Windows server and you want to run excel which is located on the server (never done this, so this is all theoretical) . I'm imagining there will be a shortcut on the workstation that will allow you to launch the application.

As far as communications concern, the workstation and server would normaly be connected via a local lan running the Ethernet II frame and TCP/IP. Now under these conditions, you would have NETBIOS bound to the TCP/IP interface. The netbios protocol is used amongst windows machines when they are in a network. Prime example is a simple share, where you can access a share on the server from your workstation. All your requests and the data are encapsulated within the Ethernet II frame and use TCP/IP to travel from one pc to another.

So, Netbios is the communication protocol which will carry all the data (the application, saves e.t.c), while TCP will be the Transport protocol which will ensure that the data gets from one computer to another.

Now because I am not exactly sure how you would do this, lets assume you launch the application from the server and it runs on the workstation. So, firstly, you double click on the icon on your desktop, this would generate a request, the application layer would grab it and format the request in a way that the other end will understand it. It then passes it to the lower layers ( I wouldnt know exactly what each layer does since I didn't program excel !) . The session layer would be respsonsible to setup the connection with the server i guess, so it would probably send a few packets in advanced to ensure the server is on 'standby' for whats the follow (the request to start the application).

The properly formatted request to start the excel application is finally passed down to the session layer and then goes to the transport layer which is where the proper ports are selected to ensure the server and workstation communicate at the correct ports and don't clash with other applications or services that might be running on other ports. The transport header is added here and then the packet goes to the network layer, where the IP header is added to the packet. The correct IP details are entered here (source IP, destination ip) and the packet is then sent to the datalink layer.

The workstations MAC address is added here and there is a CRC calculated on the entire packet.

Next is the physical layer, it will take the whole frame and transmit it over the wire to the server.

The Server receives the packet.

Now we start to go from the bottom of the OSI to the top ! So.... the physical layer receives the signals and reconstructs the Ethernet II frame and passes it to the datalink layer.

The datalink layer checks the frame to see if the frame was for its self (it checks the destination MAC address, which should match its own) and also if there are any errors. Assuming there are none, it strips the FCS, MAC and LLC header and passes the remains to the network layer. The network layer checks the source and destination IP, if they are valid, it strips the associated headers for this layer and passes it to the transport layer. The transport layer checks the ports to see if they are valid, and assuming they are, it strips its headers and passes it to the session layer.

At whole mystery starts at this layer as I already have explained, I am not 100% sure what happens. The session layer's data would be the actual request to open the application Excel. The server will interpret the request and launch the application and start sending the information to the workstation via a TCP/IP stream.

When we analyse the various protocols and compair them with the OSI model, you will always notice that very little information is provided about what happens at the upper layers. The people who program the software are able to explain that better than what a network engineer can [img]images/smiles/icon_smile.gif[/img]

I hope the above helps you solve your questions and gives you a better understanding.

Cheers,