Network Online Course
Lesson 2 - Communication Standards
B. OSI Reference Model
OSI Reference Model
- International Standards Organization
- Open Systems Interconnect Reference Model
- 7 layer model of networking
- Each layer only communicates to one above and below
- Very slow to be implemented
OSI Reference Model
For a greater understanding of data communications and networking, the Open Systems
Interconnect (OSI) seven layer reference model of networking is usually used. The model was
designed by the International Standards Organization (ISO) in order to clarify the
relationship of various network components to each other.
The purpose of the model
The OSI model is a model of how communications might work. It was designed by the
International Standards Organization with input from many other organizations including
standards committees, manufacturers and governments. The model includes all of the functions
necessary for communications and divides them by functionality into seven divisions or
layers. The purpose of this model include the following
Figure 14: The OSI Reference model of communications
- to provide general design guidelines for data communication systems.
- to divide the communication process into well-defined, functional areas (called layers),
facilitating the creation of network products and encouraging the interchangeability of
- to promote the goal of communications between different types of systems by encouraging the
development of internetworking devices.
- to make sure that all functions needed for communications are addressed and that none are
- to avoid duplication of functions thereby increasing efficiency.
- to design independence for each layer from other layers. This facilitates the
interoperability of products from different manufacturers and insures that the end-user can
substitute one vendor’s product for another.
- to produce a learning tool that can be used to understand how communications in modern
computer systems works.
Some basic rules
The OSI model has some basic rules. They include
- a layer may only communicate with the layer above and below and may not circumvent these
- a layer must be independent of any other layer.
What the layers do
The following is a discussion of each layer. Figure 14
illustrates the seven layers in their
The application layer is responsible for providing network services to users and their
applications. It is not, however, the user’s application or the operating system, such as
- It provides a user interface — a good example is the type of terminal emulation a host
requires, such as VT100 or 3270.
- It provides network services such as file transfer, mail service or terminal emulation
e.g. FTP, Telnet, SNMP.
The presentation layer deals with syntax and grammatical rules for presenting data to the
application layer (not the user). This includes encoding, decoding and otherwise converting
data. It is responsible for the following.
- Character sets — the code used to represent characters e.g. the ASCII of personal computers
vs. EBCDIC, used by IBM mainframes. A gateway has to convert character sets when a PC
communicates with an IBM mainframe.
Created by ANSI
||Used by PCs
Controlled by Uniforum
Japanese Industrial Standards Committee
||Japanese character set
|ISO-8859 Latin 1
||Used in HTML
||Used by IBM mainframes
- Compression and decompression of data — some communication protocols include compression of
data. Kermit is one such protocol.
- Encrypts and decrypts data — moving data over a network safe from prying eyes is the job of
encryption. There are many methods available for this including DES, RSA and SSL and
public/private key schemes.
- Bit order translation — numbers are sent over the network as binary numbers. The meaning of
the numbers is different if the computer starts with the leading bit first or the trailing bit.
One is the most significant bit (MSB), the other is the least significant bit (LSB).
- Byte order translation — just as a byte is made up of bits, a message is made up of bytes.
Should the most important byte be transmitted first or the least important? Intel
microprocessors transmit the least significant first (called little Endian) while Motorola
CPUs (Macintosh) transmit the most important first (called big Endian). Communication between
PCs and Macintoshes must take this into account.
- File structure — another difference between PCs and Macintoshes is their file structure.
Macintosh files are comprised of two files, called forks: the resource fork and the data fork.
PC files are single entities.
A session is an agreement to communicate between two entities and the session layer controls
the setup, termination and other mechanics of this conversation.
- It establishes and maintains connection
e.g. intent to transmit, was it successful?
- It deals with name recognition (computer name, user name) and login.
- It deals with synchronization of data transmission by placing checkpoints within data
stream so that if interrupted, the transmission can take up where it left off.
- It deals with upper layer errors in the communications process including problems with
memory, storage space, printing.
- It handles remote procedure calls (RPCs), e.g. running a program on a remote computer.
The transport layer provides extra connection services including error correction.
- It controls data flow, e.g. slows down transmission if buffers about to overflow.
- It fragments and reassembles data.
- It acknowledges successful transmission.
- It correct faulty transmission.
The network layer is primarily responsible for getting information to the correct computer on
the correct network.
- It moves information to the correct address.
- It assembles and disassembles packets.
- It determines addressing and routing.
- It determines the best path.
Data Link layer
The data link layer creates the entity (the packet or frame) which is put onto the
- It controls access to the communications channel.
- It controls the flow of data.
- It organize data into logical frames.
- It identifies specific computers on the network.
- It detects errors.
The physical layer is not the transmission media (cable) itself, but is responsible for the
specifications for the media and the electrical signal that goes on it.
Figure 15: Data flow through the OSI model
- It provides electrical and physical interface to the network.
- It specifies type of medium.
- It specifies how signals are transmitted.
Making a request
The process of moving data through the OSI model is illustrated above. It shows a request
coming from a user at a workstation to a file server. The data, for example, the request to
open a Word file, starts from the user. In fact, it starts with the Word program itself. It is
passed to each layer where it is acted upon until it finally emerges as a packet on the network
cable. It is received by the network interface card of the destination machine and then is
passed through successive layers until it emerges at the top and is passed to the program
Notice two things about this process. There may not be a distinct piece of software for each
of these layers. The functions of several layers may be combined into a single program.
Secondly, layers may be missing entirely if their functions are not needed. OSI is just a
model which is useful for learning about data communications and one which developers should
try to design for. It does not necessarily represent the real world.
Figure 16: Encapsulation
The process of moving data through successive layers is called encapsulation. Each layer takes
the frame from the layer above. This becomes the new data or “payload”. It then adds
information fields to the front of the data and possibly, an error correcting field to the
back and then passes it on to the layer below where the complete frame becomes the new data.
Because beginning and ending fields frame the data, this process is often called “framing”.
Because the data is cradled within or enclosed inside, the process is also called
At the receiving end, the process is reversed. At each layer, the framing fields for that
layer are stripped off. Error correcting information, if any, is calculated. Then the data is
passed up to the next layer and the process is repeated.
Although the process is cumbersome with a great deal of housekeeping involved, performance
can still be excellent because it operates at electronic speeds. Improving the speed of
communications is a priority for all those involved in this process, including the developers,
vendors and administrators of the systems. An important element of fine tuning, however, is
understanding the process as outlined above.
Exercise 2-1: OSI Model Quiz
Exercise a: Label the layers of the OSI model
Exercise b: Match the OSI layer with the function it provides
Mediate access to communications channel; Organize data into frames
Data representation (ASCII); Data compression and encryption
Establish and maintain connections; Deal with upper layer errors
Ensure reliability of delivered packets; Data flow and transmission errors
Specifications for physical channel; Carry signals for all higher layers
User interface; User applications
Network addressing and routing; Packet assembly and disassembly
[Top of page][On to next section]