Videcon Technical Manager Ian Farr introduces basic network theory
Take the concept of a single computer: this on its own is very useful however it is limited in what it can achieve. If we link this up to an entity such as the internet it becomes much more useful. Obviously in linking it up to the internet we come up with a number of questions such as: how do we identify the computer?

We can uniquely identify a computer by using a Mac address, coded into every ethernet device such as a network card. It is not set by the user but by the device manufacturer. The numbers used are issued from a central issuing authority, maintaining a unique address for each device. The issuing authority of the Mac address is the IEEE.

When we connect more than one computer to another we create a Local Area Network (LAN). To create the LAN we will require a number of new items. The first is a network card used to create a connection point from our computer into our network. The second is a switch, this is used to pass the packets of data around, and a cable network to link them together.

The network cable used today is typically Cat5 cable and it connects using a method referred to as 100BaseT which has a carried data speed of 100Mbps. As well as physically cabled networks, more people are opting for wireless networks. To achieve a wireless network simply install a wireless network card into the computer in place of the traditional wired LAN adaptor.

Then put into the switch a wireless receiving device such as an access point, which will convert from the wireless to wired network. The wireless network operates on a standard known as IEEE 802.11b which has a data rate at present of 11Mbps. When connecting a number of computers together you can also use a device called a hub. When it receives a data packet to one of its ports it will send that packet to all other ports on the hub. The ports that don't have the device connected will simply ignore it. A switch will take the data request given to one port and then send that information to the port that the device is connected to.

This leaves all of the other ports free and in fact on these ports you can have other concurrent connections. In comparison the data throughput of an eight-port switch could be up to four times higher than an eight-port hub. Now that we have them all linked together we need to address them uniquely. We could use the Mac address, however it is a non user-friendly format and not practical for users to use and remember, therefore we use what is known as an IP address.

As long as each address used is unique in the network then we can now identify the computers. There are two types of addresses used and these are referred to as public and private addresses. Public addresses can be accessed by everyone on the internet. Public addresses can be routed through the internet routers, allowing them to be used for public services such as web and mail. Private addresses are reserved for use in company networks and are not routed through the internet. This allows the same private address to be used in more than one private network.

If one of the computers on our network has the internet connected to it as well, then it will have an IP address on our network (private) and an IP address on the internet (public), therefore if any of the computers on our LAN wish to talk to the internet they will need to talk through this computer using its private address. Because of this path (or route) through the computer, the computer with the internet connection is referred to as our gateway (this could also be a dedicated router). The gateway is our connection point between two networks. For an example of most of the above details see the diagram.

Either the computer or the router could also be a firewall. What is a firewall? It is a device that your connection passes through with added security that allows you to stop your network being penetrated from outside.

In order to do this it will filter which requests into and out of the network are valid. When we extend our network from covering one office to covering a number of offices we have created a Wide Area Network or WAN.

TCP/IP
What is TCP/IP? Two protocols working together. Transmission Control Protocol does the verification of the data and checks that what was sent is what has been received. If not, then it will request the data be sent again. IP or Internet Protocol does the actual transport, location and addressing of the data. TCP/IP as a whole is a Unicast protocol that allows a piece of data to be transmitted from one computer to another but does not support sending of that same piece of data to many machines at once. Therefore to send the same data to three different computers it would need three separate transmissions.

A firewall is a device which your connection passes through, with added security that allows you to stop your network being penetrated

At present with IP V4 the address is a binary number, 32bits in size. This is normally represented in decimal format split up into four blocks. Because of this the numbers in each block can be in the range 0 to 255 (an example would be 192.168.0.1). There are two ways to assign an IP address, Static and Dynamic. Static addresses mean you enter into the computer the address the computer is to use to communicate.

There are many advantages to static address such as:

  • The computer is always available at the same address and because of this the data sent from it can be traced back to it.
  • Because it is always available it is especially suitable for server-based applications where the computer always needs to be locatable.

    Disadvantages:

  • Increased setup time
  • Static public addresses are more expensive
  • Because you are always available on the same address, information you have sent can be traced back to you. Also it means you are always available at the same spot for people to break into your system. When the address is assigned dynamically on your own network the computer is referred to as a DHCP (or Dynamic Host Configuration Protocol) server. When you receive a dynamic address from an ISP it is usually only referred to as a dynamic address though this is exactly the same thing as from a DHCP server. This method also has benefits and drawbacks.

    Benefits:

  • Improved IP management. IP addresses are not wasted. For example if you only had 12 connections into your network but there were a possible 15 machines that could connect, then you would only need to have 12 addresses and the DHCP server would assign the address to the connected computers as required.
  • Dynamic addresses from ISP will be cheaper as you are not paying for fixed public IP addresses.

    There is one major drawback of dynamic addresses and that is they are not suitable for server applications such as a web server or a mail server which will always need to be locatable so that people can access it or mail can be routed to it.

    With TCP/IP an important component is the subject mask, whose job is to check that the address you are sending the data to is on the same network as yourself. If the subnet mask –when applied to both the destination address and the address you are at (by means of a calculation) – comes up with the same result then it knows it is on the same network.

    If the result is different then the data is sent to the gateway because TCP/IP knows it is on a different network. This is not to say that two different networks cannot have the same subnet mask, they can, as it is the result of the sum that is important, not the subnet mask number itself.

    On the internet there are many services today that help the internet function and one of the key ones is the DNS or Domain Name Service which is used to resolve domain names.

    For example, every device on the internet that is accessible has an IP address. However these are not easy to remember so what happens is we use a domain name (e.g. www.TeleEye.com) and this address is then resolved, using the DNS server, to an IP address.

    This IP address can then be used in the normal manner in order to get from your computer to the destination. This DNS service is transparent to the user so that they do not have to do anything to use it unless they are using static IP addresses, in which case they will need to enter the address of the DNS servers into their network properties.