The TCP/IP Network: Explanations

The TCP/IP Network: Explanations

 

The communication protocols:

Before we get into it, I would like to tell everyone that this isn’t a course about networking. We will actually be vulgarising computing vocabulary so that it can be clearly explained to people with less knowledge. I will try and keep it quick and simple.

Transporting data requires physical support (in the overall scheme of things), but also logical architecture that will assure our quality of service. The exchange of information requires at least two parties that can understand each other (like what I’m trying to do with you). As well as this, the communication cant proceed unless all the other connections respect the same conventions. This can be TCP/IP, IPX/SPX or NetBEUI to use the commonly used protocols of today. These are what define your network rules.

TCP/IP:

TCP and IP were developed (in the 1970’s) by the Department of Defence with the goal of standardising all the connections between their different machines (UNIX mostly – we need to be serious after all), that were running different languages. They especially needed to unify all the potentially useable networks. This communication architecture is just like the OSI model that arrived a little later (1978). The OSI model defines 7 layers with TCP/IP having a rapid increase in use due to all the different available services:

• IP: At the network layer, to have a service without a connection.
• TCP: At the transport layer to have a solid service when connected.
• FTP: to transfer files.
• SMTP: For messaging.
• TELNET: To see what is happening on the screen (Remotely).
  

Due to this, TCP/IP became a complete protocol, which is the most widely used today.

To summarize:

IP (3^rd layer of the OSI model) is responsible for the path that data takes from one machine to another (nodes). To do this we use an address (e.g. 213.186.33.19) to identify the source and the destination. It is a global organisation (NIC) that manages these addresses, which are all unique. Each country also has to manage its addresses as they will be requested by companies etc. These demands are either managed by themselves or by internet hosts etc. To reduce the number of potential destinations, as each of us can be destinations, we use a network subnet mask which allows us to be grouped together. The mask allows us to address everyone else in the same group. These groups can be by name, which will be in a domain (DNS) and will take the form of: network.name.domain. These names are replicated in large files and are stored in domain servers (it is useful to know the addresses of these so that you can quickly find a destination). When you also address a destination (without specifying its IP address) you verify its name with the primary server. If it isn’t there then it will look in the secondary etc. By looking up the tree until it is found, you will find out its address and you will be able to communicate with it.

TCP (Layer 4 of the OSI model) is responsible for verifying the exchanged information between the client and the server. In the event there is an error, it is TCP that will resend the corrupt packets. The data is varied and the packets are of different sizes and they can be sent via different nodes in a random way, meaning that the arrival order is not always correct. Each node resends the packets if they are not for him. Otherwise the message will be put back together at its arrival. In the event there is an error the packets will be resent.

We should also mention the port number, which will identify the service (application) that is transported by the packet. These normally figure in a table for each device.

Three addresses are essential for a TCP/IP connection. Your IP address which will be used to identify you. Your subnet mask which will determine your group and finally the default router which will be the node that allows you to exit your private network and travel around the outside networks. In the case of your internet connection, you will generally not know your IP address as this will be given to you when you call your access provider with your modem. This will verify your account and grant you a temporary IP address (this will be done via DHCP), so that there are no double ups (identical IP addresses on a network can be quite catastrophic). The server which automatically allocates a dynamic IP address to you will be different then that of your private network. Your subnet mask corresponds to the range of leased addresses by your access provider and this will determine the maximum number of connected users.

I hope this has been informative and simple.