TCP/IP protocol: what it is and how it works
The TCP/IP protocol is a list of regulations designed to establish the correct transmission and receipt of bundles of information within a network. There are also many others, but TCP/IP is one of the most widely used, as it is considered reliable and almost completely error-free.
The preservation of a message until it reaches its destination is of course, one of the most important factors in accurate communication between online devices (whether wireless or cable). This family of protocols provides reliable results and loss of information bundles is kept to a minimum.
What does the TCP/IP protocol consist of?
One of the most distinctive aspects of the TCP/IP protocol is its layering system. The suite of protocols included in TCP/IP is constructed in layers according to a very specific order. This approach offers some attractive advantages, such as:
- Modularisation;
- Complex tasks can be subdivided into simpler component tasks;
- Abstraction of the problem.
It should also be noted that during communication (via TCP/IP protocol) between two interlocutors, each level forms a logical connection with its corresponding level in the destination stack.
The levels of the TCP/IP suite are normally five: application, transport, network, connection and physical. The name of the protocol derives from two very important elements present inside these levels: TCP for the transport level and IP for the network. This set of protocols is also commonly known as the TCP/IP family of protocols.
How communication via TCP/IP works
Communication between two interlocutors via the TCP/IP protocol works in a relatively straightforward way. Each Host (interlocutor) integrates the TCP/IP family of protocols, with its typical five level structure, into its own end-system.
Before the message is sent by the source host, it is packaged in layers like a matryoshka; this package is known as a bundle. The term matryoshka is used because each level of the suite generates a header and a bundle based on those of the previous level, like a ‘box’ encased in successively larger boxes.
In other words, at level 1, the message (bundle 1) and header 1 are generated. At level 2, the bundle consists of everything contained in level 1 (so bundle 1 and header 1) and a new header (header number 2). This process is repeated until it passes through all 5 levels. This procedure is known as encapsulation.
When the message is received by the destination host, the decapsulation process can begin. This is basically the opposite process and consists of the unpacking of each bundle until the original message is revealed.
As there are many protocols on each level, the packaging operation can also take place using a more complex process, known as multiplexing. In that case, the bundles received from the higher level are more than one and can be directed to the correct protocol in the next level using the corresponding headers.
The latter, as well as containing specific information for the next level, also contain the destination address of the bundle once it reaches the receiving host (where it will undergo the extraction procedure known as demultiplexing).
Each TCP/IP protocol level has clearly-defined addresses: the connection level, for example, uses MAC addresses, the network level uses IP addresses, the transposing level uses Ports, while the application level uses the Name (http://nomesito). Thanks to this differentiation there is no risk of a bundle ending up in the wrong level when it reaches its destination.
Translated by Joanne Beckwith
