The Layers of the TCPIP Model

The Layers of the TCP/IP Model

The TCP/IP model was developed by the U.S. Department of Defense (DoD) in the late 1970s. A number of standards have been defined for this model and are listed in Request For Comments (RFC) documents. These documents describe the architecture and use of the various protocols and layers. The RFCs are listed and can be retrieved on the Web at http://www.ietf.org/rfc.html. For example, RFCs 0760 and 0761 describe the DoD standard Internet Protocol (IP) and standard Transmission Control Protocol (TCP), respectively.

The TCP/IP model is described as a four-layer model that functions on a common hardware platform. The four layers are Application, Transport, Internet, and Network Interface. This model has also been commonly described as a five-layer model, which treats the hardware layer as a separate layer. This discrepancy arises as a result of interpretation, depending on whether the Physical layer is treated as part of the Network Interface layer or not.

Sun describes the TCP/IP model as a four-layer model on a common hardware platform. However, in the exam objectives for the Solaris 8 Network Administrator exam, Sun wrote "Identify the purpose of each layer in the TCP/IP five layer model." You should read any question relating to the number of layers very carefully . For example, if you are asked to list the 4 layers of the TCP/IP model, then the answer would be "Application," "Transport," "Internet," and "Network Interface." If you are asked to list the 5 layers of the TCP/IP model, then include "Hardware" in your choice of answers.

This book treats the TCP/IP model as a four-layer model because the published Sun Microsystems objective for this exam mentions only four layers ”the inference is that the Hardware layer is integrated into the Network Interface layer. Figure 2.1 shows the hierarchical nature of the TCP/IP model layers as well as each layer's position in the hierarchy.

Figure 2.1. The TCP/IP model layer hierarchy.

Network Interface Layer

As you can see from Figure 2.1, the Network Interface layer is the lowest in the TCP/IP hierarchy and provides the communication service to the Internet layer. The Network Interface layer prepares data packets received from the Internet layer into suitably sized packets for transmission across the physical network. Also, it receives packets from its peer layer on another host and passes them upward to the Internet layer. Addressing at this layer is based on the MAC address. Specifically, this layer does the following:

• Handles the transmission of data across the physical network.

• Assembles data for transmission into Ethernet frames. (Ethernet frames are discussed in Chapter 1.)

• Detects errors through the use of the Cyclic Redundancy Check (CRC) field in the Ethernet frame.

Internet Layer

The Internet layer resides above the Network Interface layer, is responsible for the routing of data to the destination host, and tries to ensure that the most efficient route is taken. IP addresses are used at this layer, unlike the Network Interface layer, where MAC addresses are used. This layer uses a routing table to find the next host on the network that is on the route to the destination. This might be the actual destination node, or the nearest gateway. Data packets at this layer are called IP packets . Chapter 4, "Fundamental IPv4," and Chapter 7, "Basic IPv6," discuss the IP protocol and addressing in more detail.

Transport Layer

The Transport layer is responsible for the end-to-end delivery of the data through the use of

• Sequencing of packets so that receipt of all packets can be verified .

• Acknowledgements so that the sending host can verify that each packet has arrived at its destination.

• Flow control so that network congestion can be handled effectively.

The Transport layer is located above the Internet layer and provides a more reliable service. The Network Interface layer, for example, might encounter a CRC error, indicating that a data packet was corrupted ”this layer would flag the error, but wouldn't request a retransmission of the failed packet; the Transport layer would handle this. This layer is discussed in more detail in Chapter 9, "TCP and UDP."

Application Layer

The Application layer is the top layer of the TCP/IP model and interacts with the running processes that make use of network communications. Examples of network communication processes are Telnet, File Transfer Protocol (FTP), Dynamic Host Configuration Protocol (DHCP), Domain Name System (DNS), Network File System (NFS), email, and of course, the Web browser.

Specifically, this layer does the following:

• Handles the data format ” For example, when a Unix system is communicating with an IBM system, the Unix system encodes in ASCII (American Standard Code for Information Interchange), whereas IBM uses EBCDIC (Extended Binary Coded Decimal Interchange Code). This layer handles the data conversion.

• Manages application connections (sessions) ” This layer handles how applications interoperate between the two connecting hosts , an example being when an NFS client connects to the NFS server to mount a file system, using Remote Procedure Calls (RPC), or when a user accesses a page on a Web server when using a Web browser.

• Handles how the end user wants the data presented ” In addition to handling the formatting of data, this layer also handles how data is displayed, such as in the formatting of email messages.