Network communication relies upon the interaction of many different protocols. These protocols are designed to implement a particular set of rules and conventions governing particular aspects of how devices in a network communicate.
Network protocols are arranged in layers, with each layer providing a specialised service for the layer above. Because each layer consists of standardised protocols that perform functions to an agreed set of rules, it is possible to modify individual protocols within a layer without having to change protocols in other layers.
Rules: agree to speak a common language, take turns speaking, not to interrupt when others are speaking.
Physical: face-to-face communication requires speakers to be close enough to hear each other.
In this example, the conversation has been divided into three layers. The physical layer considers the fundamental requirements of exchanging information between two people using speech. This includes considering the distance between the speakers, which will affect how loud each person needs to talk in order for the other to hear. Speech is a movement of air, which is translated as sound waves. This means that the conversation could take place in a ‘near’ vacuum like outer space.
The rules layer defines how the speakers will talk. To exchange information, a common language would need to be chosen. Once the language is agreed, how is the conversation to be managed? Taking turns is a reasonable way to converse with another person and interrupting would be considered rude. But what happens if one person misses something and wishes it to be repeated? Do they ask for it to be repeated verbally, or do they use an agreed signal such as raising a hand?
The content layer considers the actual message exchanged between the two speakers, which in a conversation may be an item of news, plans for a holiday or just gossip.
By dividing the conversation into layers, we have decided upon the particular functions that are required by each layer to support communication, and in our simple layered model, it should be possible to change some aspects within one layer without affecting the others. For example, if we now decide to have a telephone conversation, how does this change the requirements of our physical layer?
Networking protocols can be layered in a similar manner to the voice conversation in the example above, but has a wider range of considerations and uses different terms:
These protocols operate on your computing devices, and allow the programs you use to access the data network via the single network interface card (NIC) that is fitted to most devices. Imagine how difficult this task would be for your device if you used multiple programs that required network access – for example, surfing the World Wide Web (WWW), downloading a file and sending an email.
All these different programs expect to be able to simultaneously access your NIC in order to send and receive data. The purpose of the TCP/IP reference model is to allow the design and creation of protocols that can support the network requirements of various installed programs competing for the attention of a single NIC.
Each layer of the model defines the function of the protocols that operate within it:
Application: provides the interface between the other layers and the operating system software on your computing device.
Transport: manages application layer data and prepares it for transmission by segmenting it into easily manageable blocks. It is also responsible for allowing multiple programs to simultaneously access the NIC, and managing the reliable delivery of data if required.
Network access: manages the delivery of data across local network media, regardless of the physical media type.
Groups of protocols that are created in accordance with a protocol model are referred to as a protocol suite. The TCP/IP protocol suite operates in accordance with the TCP/IP reference model, and it is the predominant suite used within the Internet.