Definitions of OSI Model

OSI Model

OSI model. Reference framework for the definition of communication systems interconnection architectures. According to abbreviationfinder, it is a functional guideline for communications tasks and therefore does not specify a communication standard for such tasks. However, many standards and protocols comply with the guidelines of the OSI Model.


During the 1960s and 1970s many networking technologies were created. Each based on a specific hardware design. These systems were built in one piece; what we could call a monolithic architecture. This means that the designers had to take care of all the elements involved in the process.

We can assume that these elements form a transmission chain that has several parts: the physical connection devices, the software and hardware protocols used in the communication, the application programs that carry out the communication, and the man-machine interface that allow the human use the network.

This model, which considers the chain as a monolithic whole, is impractical, since the smallest change can imply altering all its elements. The original Internet design of the US Department of Defense had a four-layer scheme. Although it dates back to the 70s, it is more or less the one that is still used: Physical or Network Access Layer (“Network Access Layer”).

Responsible for sending information about the hardware system used in each case. A different protocol is used depending on the type of physical network. Network Layer also called Internet Layer (“Internet Layer”). It is responsible for sending the data through the different physical networks that can connect a source machine with the destination of the information.

Transmission protocols such as IP are closely associated with this layer. Transport layer (“Host-to-Host Layer”). Controls the establishment and termination of the connection; data flow control; retransmission of lost data, and other details of the transmission between two systems. The most important protocols at this level are TCP and UDP (mutually exclusive). Application layer (“Application layer”).

Made up of the protocols that directly serve the user programs; Browser, E-mail, FTP, TELNET, etc. Responding to the prevailing general theory in the world of computing, of designing hardware by modules and software by layers, in 1978 the ISO organization (International Standards Organization, proposed a communications model for networks which they called “The reference model of Open Systems Interconnection”, generally known as the OSI model.

Its philosophy is based on breaking down the functionality of the transmission chain into various modules, whose interface with the adjacent ones is standardized. This design philosophy has a double advantage: changing one module does not necessarily affect the entire chain. In addition, there may be a certain interoperability between various products and hardware/software manufacturers, since the limits and interfaces are perfectly defined. This means, for example, that two different communication software can use the same physical means of communication.

Main components

The OSI model has two main components:

  1. A network model, called the basic reference model (“Basic Reference Model”) or service layer (“Server-layer”).
  2. A series of concrete protocols.

The network model, although inspired by the Internet has no more similarities with that. It is based on a model of seven layers, while the primitive Internet was based on 4. Currently, all developments are based on this model of 7 levels, which are the following:

  • Physical
  • Link
  • network
  • Of transport
  • Session
  • of presentation
  • of application

Each level performs a specific function, and is separated from the adjacent ones by known interfaces, without being concerned by any other aspect of the total communication. Note that this model, to which absolutely every book dealing with networks refers, is only a conceptual abstraction, to which physical reality (more or less) conforms. However, the consequences of the application of this point of view are very useful. The underlying concept is the same that allows us to put address and return information on a letter; frank it and deposit it in the mailbox of the postal service without worrying about anything else.

The letter follows a series of processes within the postal service without each one having to worry about those that have preceded or will follow. Finally, the letter is delivered to the recipient in his mailbox, who only has to read it. Generally, the devices used in networks limit their operation to one or more of these levels. For example, a concentrator (” Hub “) that amplifies and retransmits the signal through all its ports, is operating exclusively at layer 1, while that a switch (” Switch “) operates at layers 1 and 2; a router (” Router “) operates at layers 1, 2, and 3. Finally, a user workstation typically handles layers 5, 6, and 7.

As far as the software is concerned, it should be noted that each layer uses a specific protocol to communicate with the adjacent layers, and that it adds certain additional information (” Protocol Header “) to the header (” Header “) of the packet.

The layers of the OSI model

The schematic description of the various layers that make up this model is as follows:

  • Physical layer-1- (“Physical layer”); It is responsible for transmitting the bits of information through the line or medium used for transmission. It deals with the physical properties and electrical characteristics of the various components; of the transmission speed, if it is uni or bidirectional (simplex, duplex or flull-duplex). Also of mechanical aspects of the connections and terminals, including the interpretation of electrical signals. As a summary of the tasks of this layer, we can say that it is responsible for transforming a binary information packet (” Frame “) into a succession of impulses appropriate to the physical medium used in the transmission.

These impulses can be electrical (wired transmission); electromagnetic (Wireless transmission) or light (optical transmission). When it works in reception mode, the work is inverse; It is responsible for transforming these impulses into binary data packets that will be delivered to the link layer (see below). For example: this level defines the measurements of the Ethernet coaxial cable and the BNC connectors used. Other examples of standards related to this layer are RS-232 (H2.5.1) for serial communications and X.21

  • Linklayer -2- (“Data Link layer”). It can be said that this layer transfers the messages to/from the physical layer to the network layer (which we will see next). It specifies how the data is organized when it is transmitted in a particular medium. This PE layer defines the frames (“Frames”), the addresses and the checksums (” Checksum “) of the Ethernet packets.

In addition to local addressing, it deals with the detection and control of errors that occur in the physical layer, the control of access to said layer and the integrity of the data and reliability of the transmission.

For this, it groups the information to be transmitted in blocks (“Frames”), and includes a checksum to each one that will allow the receiver to verify its integrity. The received datagrams are checked by the receiver. If any datagram has been corrupted, a control message is sent to the sender requesting its re-sending. The PPP protocol [1] is an example of this layer.

The link layer can be considered divided into two sublayers:

  1. Logical Link Control LLC (“Logical Link Control”) defines the way data is transferred over the physical medium, providing service to higher layers.
  2. MAC medium access control (“Medium Access Control”). This sublayer acts as a controller for the underlying hardware (the network adapter). In fact, the network card driver is sometimes called the “MAC driver”, and the address The physical address contained in the hardware of the card is known as the MAC address (“MAC address” H12.4).

Its main task (which gives it the name -access control-) consists of arbitrating the use of the physical medium to make it easier for several teams to compete simultaneously for the use of the same means of transport.

The CSMA/CD (“Carrier Sense Multiple Access with Collision Detection”) mechanism used in Ethernet (H12.4) is a typical example of this sublayer.

The IEEE 802.11 standard

The IEEE 802.11 standard defines the use of the two lower levels of the OSI architecture (physical and data link layers), specifying their rules of operation in a WLAN. The 802.x branch protocols define the technology of local area networks and metropolitan area networks.

802.11i was ratified on June 24, 2004 to address the issue of wireless network security. It is based on the TKIP encryption algorithm, like WPE, but also supports the much more secure AES (Advanced Encryption Standard).

  • Networklayer -3- (“Network layer”). This layer deals with the transmission of datagrams (packets) and routing each one in the right direction (“Routing”), a task that can be complicated in large networks like the Internet, but it doesn’t deal with errors or packet loss at all. For example, it defines the address and route structure of the Internet.

Two types of packets are used at this level: data packets and route update packets. As a consequence, this layer can be considered subdivided into two:

  • Responsible for encapsulating the data to be transmitted (user). Uses data packets. In this category is the IP protocol (“Internet Protocol” A3.1).
  • Switching (“Switching”): This part is responsible for exchanging specific connectivity information of the network (its activity is rarely perceived by the user).

Routers are devices that work at this level and benefit from these route update packets. In this category is the ICMP protocol (“Internet Control Message Protocol” A3.7), responsible for generating messages when transmission errors occur and a special echo mode that can be checked by PING [3].

The most frequently used protocols in this layer are two: X.25 and IP. Webography: OSI: The Network Layer This Cisco Systems page provides an authoritative summary of network layer services and protocols.

  • Transportlayer -4- (“Transport layer”). This layer is responsible for guaranteeing the reliability of the service, describes the quality and nature of data delivery. PE This layer defines when and how retransmission should be used to ensure its arrival. To do this, it divides the message received from the session layer into pieces (datagrams), numbers them consecutively and delivers them to the network layer for sending.

During reception, if the Network layer uses the IP protocol, the Transport layer is responsible for reordering the received packets out of sequence. It can also work in reverse by multiplexing a transport connection between various data connections. This allows data from various applications to share the same flow to the network layer.

A typical example of protocol used in this layer is TCP (“Transport Control Protocol” A3.1), which with its IP counterpart of the Network layer, configure the TCP/IP suite used in the Internet, although there are others such as UDP (” Universal Datagram Protocol”) a transport layer also used on the Internet by some application programs.

  • Session Layer-5- (“Session Layer”). It is an extension of the transport layer that offers dialog control and synchronization, although in reality there are few applications that make use of it. For example, Internet communications do not use it. Note: Some authors indicate that the session layer is merely a theoretical consideration of the model authors with absolutely no known practical utility.
  • Presentation Layer -6- (“Presentation layer”). This layer deals with the semantic aspects of communication (describes the syntax of the data to be transmitted), establishing the necessary arrangements so that they can communicate with machines that use different internal representations to data.PE describes how floating point numbers can be transferred between computers using different mathematical formats.

This layer is a good candidate for implementing cryptography applications. In theory this layer “presents” the data to the application layer by taking the received data and transforming it into formats such as text, images and sound. As we will see below, in reality this layer may be absent, since few applications make use of it. her. Note: Something similar to the previous one happens with this layer.

In theory client and server had to negotiate the format to use, and this function, and the corresponding data formatting, would be the object of this layer. However, this, which made some sense in the 1970s, when much networking was concerned with inputting and outputting data to large computers using “Dumb” terminals of various kinds (using slightly different control codes). different) doesn’t make much sense anymore.

Currently the panorama has changed; There is only one option for the data format, despite which the OSI protocol still negotiates an encoding scheme (the only one available). On the Internet, the only service that uses this layer is TELNET, which is precisely a service for accessing servers from remote terminals. In this case, the presentation layer is the one in charge of configuring the terminal to connect to a server with particular characteristics.

  • Application Layer-7- (“Application layer”). This layer describes how the application programs (browsers, mail clients, remote terminals, file transfer, etc.) do their work. For example, this layer implements the operation with system files. On the one hand they interact with the presentation layer; on the other, they represent the interface with the user, giving him the information and receiving the commands that direct the communication. Examples of protocols used by programs in this layer are HTTP, SMTP, POP, IMAP, Modbus Communication Protocol, etc.

Operation of the network layer in the OSI model

The network layer provides its services to the transport layer, being a complex layer that provides connectivity and selection of the best path for communication between machines that may be located in geographically different networks.

It is responsible for the functions of switching and routing information (logical addressing), providing the necessary procedures for the exchange of data between the source and the destination, so it is necessary to know the topology of the network (the way in which nodes are interconnected), in order to determine the most appropriate route.

Its main functions are:

  • Divide transport layer messages (segments) into more complex units, called packets, to which it assigns the logical addresses of the computers that are communicating.
  • Know the topology of the network and handle the case in which the source machine and the destination machine are in different networks. Route information through the network based on packet addresses, determining switching and routing methods through intermediate devices (routers).
  • Send packets from node to node using a virtual circuit or datagrams. Assemble the packages on the destination computer. This layer is where the routers work, devices in charge of routing or directing the data packets from the origin to the destination through the best possible route between them.

How IP works within the OSI model

The IP protocol is the fundamental basis of the Internet. It makes it possible to send data from the source to the destination. The transport layer breaks the data stream into datagrams. During transmission, a datagram can be broken into fragments that are reassembled at the destination.

IP packets:

  • It’s 4. Allow updates.
  • The length of the header in 32-bit words. The maximum value is 15, or 60 bytes.
  • Type of service. Determines if the sending and speed of the data is reliable. Not used.
  • Total length. Up to a maximum of 65,535 bytes. I
  • To determine which datagram a fragment belongs to.
  • DF (Don’t Fragment). The destination cannot mount the datagram again.
  • MF (More Fragments). Not set in the last snippet.
  • Fragment offset. Which part of the datagram this fragment belongs to. The elementary chunk size is 8 bytes. Time of life. Each hop is decremented.
  • Transport protocol on which the datagram should be based. Options include strict routing (the entire path is specified), loose routing (only a few routers are specified in the path), and route recording.

The technical operation in which data is transmitted through the network can be divided into two discrete, systematic steps. Each step performs certain actions that cannot be performed in another step. Each step includes its own rules and procedures, or protocol. The steps of the protocol must be carried out in an appropriate order and must be the same on each of the computers on the network. On the source computer, these steps have to be carried out from the top down.

On the target computer, these steps have to be carried out from the bottom up. The origin team. The protocols on the source computer:

1.They are divided into smaller sections, called packages.

2.Information about the IP address is added to the packets, so that the destination computer can determine if the data belongs to it. 3.Prepare the data to be transmitted through the NIC and sent over the network cable.

The destination computer: The protocols on the destination computer consist of the same series of steps, but in reverse.

1.It takes the data packets from the cable and feeds them into the computer through the NIC.

2.Extracts all the information transmitted from the data packets, eliminating the information added by the source equipment.

3.Copy the packet data into a buffer to reorganize it and send it to the application. The source and destination computers need to perform each step in the same way so that the data has the same structure when received as when it was sent.

Processing of TCP/IP packets in the OSI model

Protocols like TCP/IP determine how computers communicate with each other over networks like the Internet. These protocols work together, sitting on top of each other in what is commonly known as a protocol stack. Each protocol stack is designed to perform a special purpose on the sending and receiving computers.

The TCP stack combines the application, presentation, and session stacks into one, also called the application stack. In this process, the characteristics of the packaging that takes place to transmit data are given: The TCP application stack formats the data that is being sent so that the lower stack, the transport stack, can send it. The TCP application stack performs the equivalent operations that are performed by the top three OSI stacks: application, presentation, and session.

The next stack is the transport stack, which is responsible for data transfer, and ensures that the data sent and received are in fact the same, in other words, that no errors have occurred during the sending of the data. TCP divides the data it gets from the application stack into segments. Add a header containing information that will be used when the data is received to ensure that it has not been altered en route, and that the segments can be correctly combined back into their original form.

The third stack prepares the data for delivery by inserting it into IP datagrams, and determining the exact Internet address for it. The IP protocol works in the Internet stack, also called the network stack. It places an IP container with a header on each segment. The IP header includes information such as the IP address of the sending and receiving computers, the length of the datagram, and the order of its sequence. The sequential order is added because the datagram could possibly exceed the allowed network packet size, and thus would need to be broken into smaller packets. Including the sequential order would allow them to recombine properly.

Main reference

One of the most pressing needs of a communications system is the establishment of standards, without them only equipment from the same manufacturer and using the same technology could communicate with each other. The connection between electronic equipment has been gradually standardized, the OSI Model is the main reference for network communications.

Although other models exist, most network vendors now tie their products to the OSI model, especially when they want to teach users how to use their products. Manufacturers consider it to be the best tool available for teaching how to send and receive data over a network.

The OSI reference model allows users to see the network functions that occur at each layer. It is an understandable model for users. In addition, in the work the IP was defined and explained both in its version 4 and in its new version, IP version 6. Understanding that the need for the creation of the new version lies in the depletion of the addresses of the previous IP. The OSI model was explained and emphasis was placed on layer 3, because the Internet protocol, that is, the IP protocol, works or operates in this layer. Communications are established in this layer and it determines the path that the data will take.

OSI Model