One of the other interesting concepts defined by the OSI model is that each layer assumes it is communicating with an identical layer on another computer. For example, if you have two computers on a network, when you connect to another computer from your computer, you establish a session. The session layer on your computer then assumes it is communicating directly with the session layer on the other computer, as shown in Figure 8.2. In reality, the data requests move up and down the various layers on each computer, but logically, these individual modules assume they are communicating with their counterparts on the other computer. As each layer moves data up or down the layers on your computer, the data is encapsulated in a wrapper. This wrapper is removed when the data reaches the same layer on the destination computer. Now, let’s take a closer look at the various layers of the OSI model and see exactly what they do.

•  Application Layer—The application layer is the beginning of the OSI model and is where your applications begin to communicate across the network. This is the interface that your applications use to access network resources. For example, Windows NT Explorer and Microsoft Exchange are instances of applications that utilize network resources and interface with the application layer. If you are using Microsoft Word for Windows and save your file on a networked drive, this too is an example of an application interfacing with the OSI model’s application layer.

•  Presentation Layer—The presentation layer takes the data passed to it from the application layer and to an intermediate format. Then, it compresses the data and encrypts it, as well (if required), before passing it to the session layer. On the way back up from the session layer, the presentation layer reverses the process and converts it from an intermediate format to a format the application layer can use, and decompresses and decrypts it, as well (if required).

•  Session Layer—The session layer is responsible for establishing and terminating a communication linkage between two computers. This layer also regulates which computer may transmit data and how long this one-way transmission may occur before the other computer gets a chance to transmit data.

•  Transport Layer—The transport layer ensures that the data is reliably sent across the network. It supplies the error recovery and recognition features. When data is transmitted across the network, the receiving transport layer will send an acknowledgment to the transmitting transport layer to assure that the data was received. If no acknowledgment is received, the data will be retransmitted. The transport layer is also responsible for repackaging messages. Repackaging is accomplished by breaking down large messages into smaller messages (as required by the lower layer) or, alternatively, building a large message from a series of smaller messages (as provided by a lower layer).

•  Network Layer—The network layer’s primary purpose is moving data across the network and determining the best route to accomplish this task. It also performs logical-to-physical address translation. In addition, the network layer acts as a traffic cop and manages the switching, routing, and congestion control of data packets on the network.
The network layer also converts data into a network packet and places a wrapper around the data. As part of this conversion process, an Error Correction Code (ECC) is inserted. If the network packet is too large to handle for the data link layer, the network layer will rebuild the single, large packet into several smaller packets that can be handled by the data link layer. On the way back up the layers, this process is reversed, and smaller packets are reassembled into larger packets before being passed to the transport layer.

•  Data Link Layer—The data link layer takes the network packet provided by the network layer and converts this into data frames. Basically, a frame is another wrapper around data that may be broken down into smaller pieces and includes additional network addresses. Data frames are network dependent, so if a data frame is sent to a NetWare server, it may differ from a data frame sent to a Windows NT server, even though they contain the same data. Like the network layer, the data link also includes additional error correction codes to ensure reliable frame transmission and reception. If the ECC code does not match, the data frame is discarded. This prevents upper layers from receiving bad data. If a frame is discarded, not received, or unacknowledged, it will be retransmitted from the source data link layer.
The data link layer is further subdivided by some implementations into the logical link control (LLC) and media access control (MAC) layers. In this model, the LLC layer provides access to the services, while the MAC layer implements addressing and error correction.

•  Physical Layer—The physical layer is responsible for moving data frames supplied by the data link layer across the network cables. The physical layer performs this task much as a serial port does. Data is received in small fragments and is sent across the cable one bit at a time. This includes the encoding and synchronization of the data bits that are sent across the cable. The physical layer also defines the characteristics of the network media. This includes how the data is sent, the network cabling, how the cable is attached, and even electrical characteristics of the cable.

Figure 8.2  Two computers establishing a session in relation to the seven layers of the OSI network model.

Now, let’s take a look at the Windows NT Network model.