[InnerPeace]

TCP/IP model


This page discusses the TCP/IP reference model, which is the historical and technical standard of the Internet.

The U.S. Department of Defense (DoD) created the TCP/IP reference model, because it wanted to design a network that could survive any conditions, including a nuclear war. In a world connected by different types of communication media such as copper wires, microwaves, optical fibers and satellite links, the DoD wanted transmission of packets every time and under any conditions. This very difficult design problem brought about the creation of the TCP/IP model.

Unlike the proprietary networking technologies mentioned earlier, TCP/IP was developed as an open standard. This meant that anyone was free to use TCP/IP. This helped speed up the development of TCP/IP as a standard.

The TCP/IP model has the following four layers:

Application layer
Transport layer
Internet layer
Network access layer

Although some of the layers in the TCP/IP model have the same name as layers in the OSI model, the layers of the two models do not correspond exactly. Most notably, the application layer has different functions in each model.

The designers of TCP/IP felt that the application layer should include the OSI session and presentation layer details. They created an application layer that handles issues of representation, encoding, and dialog control.

The transport layer deals with the quality of service issues of reliability, flow control, and error correction. One of its protocols, the transmission control protocol (TCP), provides excellent and flexible ways to create reliable, well-flowing, low-error network communications.

TCP is a connection-oriented protocol. It maintains a dialogue between source and destination while packaging application layer information into units called segments. Connection-oriented does not mean that a circuit exists between the communicating computers. It does mean that Layer 4 segments travel back and forth between two hosts to acknowledge the connection exists logically for some period.

The purpose of the Internet layer is to divide TCP segments into packets and send them from any network. The packets arrive at the destination network independent of the path they took to get there. The specific protocol that governs this layer is called the Internet Protocol (IP). Best path determination and packet switching occur at this layer.

The relationship between IP and TCP is an important one. IP can be thought to point the way for the packets, while TCP provides a reliable transport.

The name of the network access layer is very broad and somewhat confusing. It is also known as the host-to-network layer. This layer is concerned with all of the components, both physical and logical, that are required to make a physical link. It includes the networking technology details, including all the details in the OSI physical and data link layers.

Figure illustrates some of the common protocols specified by the TCP/IP reference model layers. Some of the most commonly used application layer protocols include the following:

File Transfer Protocol (FTP)
Hypertext Transfer Protocol (HTTP)
Simple Mail Transfer Protocol (SMTP)
Domain Name System (DNS)
Trivial File Transfer Protocol (TFTP)

The common transport layer protocols include:

Transport Control Protocol (TCP)
User Datagram Protocol (UDP)

The primary protocol of the Internet layer is:

Internet Protocol (IP)


The network access layer refers to any particular technology used on a specific network.

Regardless of which network application services are provided and which transport protocol is used, there is only one Internet protocol, IP. This is a deliberate design decision. IP serves as a universal protocol that allows any computer anywhere to communicate at any time.

A comparison of the OSI model and the TCP/IP model will point out some similarities and differences.

Similarities include:

Both have layers.
Both have application layers, though they include very different services.
Both have comparable transport and network layers.
Both models need to be known by networking professionals.
Both assume packets are switched. This means that individual packets may take different paths to reach the same destination. This is contrasted with circuit-switched networks where all the packets take the same path.

Differences include:

TCP/IP combines the presentation and session layer issues into its application layer.
TCP/IP combines the OSI data link and physical layers into the network access layer.
TCP/IP appears simpler because it has fewer layers.
TCP/IP protocols are the standards around which the Internet developed, so the TCP/IP model gains credibility just because of its protocols. In contrast, networks are not usually built on the OSI protocol, even though the OSI model is used as a guide.

Although TCP/IP protocols are the standards with which the Internet has grown, this curriculum will use the OSI model for the following reasons:

It is a generic, protocol-independent standard.
It has more details, which make it more helpful for teaching and learning.
It has more details, which can be helpful when troubleshooting.

Networking professionals differ in their opinions on which model to use. Due to the nature of the industry it is necessary to become familiar with both. Both the OSI and TCP/IP models will be referred to throughout the curriculum. The focus will be on the following:

TCP as an OSI Layer 4 protocol
IP as an OSI Layer 3 protocol
Ethernet as a Layer 2 and Layer 1 technology

Remember that there is a difference between a model and an actual protocol that is used in networking. The OSI model will be used to describe TCP/IP protocols.