The OSI model by IBM is the biggest boon that systems developer can have had for their karma in the past life.

What exactly is the OSI model?

As I had mentioned it is actually a layered stack, with 7 layers. Each of these 7 layers are implemented by different protocols for each layer. But keep in mind, its not necessary that all these layers have to be used, when trying to device a system. A system can also make away with any of the 7 layers like say the Session Layer or the Application Layer.

Consider a Router for example. A Router is a device, which works in the Network Layer of the OSI Model. Its main function is to route packets that arrive at its ports. Router takes a routing decision for each of the packet that arrives at its port, on the basis of Routing Tables.

Routing Tables are somewhat similar to the Time Table we follow while attending our courses in college. Depending on the subject scheduled in the timetable for a particular day, we decide whether to attend the lecture or not. In a similar way the Router looks at the destination address (which is the IP address) of the packet that is to be routed.

Now remember that an IP address is made up of a Host ID and a Network ID (I will cover IP addresses in depth in the future articles). The Router looks at the Network ID and compares it with the other Network ID’s present in its Routing Tables. If the particular Network ID is found to be present in the Router’s Table, the packet is routed to the appropriate destination successfully. If the Network ID of the packet is not present in the Router’s Routing Table, then the router forwards the packet to the next nearest router and this process continues till the packet reaches its intended destination.

This is very similar to what you do when you are asked where someone stays, while walking on the road. You are shown the address by the person, who wants to reach a place (Consider this Address as the Destination IP address). Now what you do is directly look at the name of the street. This is because the name of the building is not something you will be immediately familiar with, as a building with the same name will surely exist in thousand different streets.

If you are very well aware of the geography of the place, then by looking at the Street (Network ID) you tell the person where the street is. ( similar to a Router finding that the Network ID field of the IP address of the packet is present in its Routing Table). If you don’t know the street or the address, you tell the person to ask some one else ( that is you forward him to the next nearest Router).

OSI model doesn’t say you do things this way. It just tells you “ Hey keep in mind that these things should be covered while designing a feasible network solution “.

TCP/IP Model

The TCP/ IP model superseded the OSI model. This is the model that is currently most widely used. The various layers in the TCP/IP model are :

5) Application
4) Transport
3) Internet work
2) Data Link
1) Physical

Note that the Sessions and Application Layer present in the OSI Model are absent in the TCP/IP Model.

In the 1990’s, TCP/IP has become firmly established as the dominant commercial architecture and as the protocol suite upon which the bulk of new protocol development is to be done.

There are a number of reasons for the success of the TCP/IP model over the OSI model:

1. Internet is built on the foundation of the TCP/IP suite. The tentacles of the Internet and the World Wide Web have spread throughout the world and that is the main reason for the success of TCP/IP model over the OSI model.
2. TCP/IP protocols were initially researched under a project in the Department of Defense (DOD). DOD was committed to international standards and most of its operational requirements couldn’t be met by the OSI model. So it started to develop the TCP/IP. Since the DOD is the largest consumer of software products in the world, the vendors were encouraged to develop TCP/IP based products.

In the next article I will cover more about the various layers in the OSI Model and the TCP/IP Model.