comp.org.uk

Networking | Programming | Security | Linux | Computer Science | About

TCP/IP Layers Explained

The TCI/IP model contains four layers:

Layers use protocols. A protocol is an agreement on how something is done. In networking, a protocol defines how the information that is transmitted from one computer to another is structured. Some protocols are special function protocols, and some are networking protocols. The function of a protocol is largely determined by the layer of the TCP/IP model at which it operates.

Networking protocols provide transport services to the special purpose protocols. They also define the rules of communication between devices. In this respect, networking protocols are like languages. The devices must share at least one common language. TCP/IP is a networking protocol and is composed of a set of subprotocols that form what is called the TCP/IP suite.

The protocols in the TCP/IP suite are as follows:

The Application Layer

The Application layer is the user applications’ interface to the networking process that is facilitated by TCP/IP. When a user attempts to access anything, the computer has to decide whether the object is located locally on the hard drive or is somewhere out on the network. If the computer determines that the network is required, the Application layer begins the process of creating the package (or series of packages, in most cases) that will be used to request the object or information from the remote device, or alternatively to transfer requested information to another device. The name used to describe this package will change as it is transferred from one layer to another.

The information that is contained in the package that will be handed down to the Transport layer includes the data itself (for example, a web page or a docu- ment) and all other information required to establish the session between the source and destination service or application and to manage the format or presentation of the data. In doing this, the Transport layer on the source device uses port numbers to identify protocols and services. These port numbers are used to communicate this information to the Transport layer on the destination device. When routers make decisions about the handling of packets based on port numbers, it is said that they are operating at the Transport layer because the routers are using information that is placed in the packet at that layer.

The Transport Layer

The parts of the TCP/IP suite dedicated to using the information provided by the port numbers are located at the Transport layer. UDP and TCP are the two subprotocols of TCP/IP that operate on this layer. The protocol selected at this layer will be a function of the transmission type (unicast, multicast, or broadcast) and the port number specified between the Application layer and the Transport layer. The protocol and port number required are added to the package, and the package is handed down to the Internet layer.

The Internet Layer

At the Internet layer, the logical addresses of the source and destination devices are determined and placed on the package. This information, in the form of IP addresses, is used by the devices on the network that can operate at the Internet layer in the process of routing the package to its final destination. As you may have surmised, this is why routers are Internet layer devices. They use information placed in the package at the Internet layer (IP addresses) to make routing decisions. There are two more protocols operating at the Internet layer: ICMP and IGMP.

The Network Access Layer

At the Network Access layer (also referred to as the Network Interface or Link layer), the source and destination physical addresses are put on the front of the package in a part called the header. Information used to perform a frame check sequence on the message is placed at the back of the package in a section called the trailer.

Finally, the package is converted to ones and zeros in the format required by the physical medium in use. If the computer is connected with a cable, the format will be electrical impulses. If the computer is connected to a wireless network, the format will be radio waves. If the computer has a fiber-optic connection, the format will be light waves. This final operation is performed by the network adaptor in the computer, which is why you must have an Ethernet (wired), wireless, or fiber-optic network adaptor, respectively, in a computer to operate on those network types, so that the bit patterns (the ones and zeros) can be sent in the required format.


Published on Thu 09 February 2012 by Darren Hobson in Networking with tag(s): tcpip