NSI Logo

Computers located here
USAFranceUSSRGreeceUK
HOW TO SET-UP A LOCAL AREA NETWORK


The easiest way to set-up your network is to click this button LET NSI DO IT FOR ME! and let us do it for you - BUT if you would like to try and do it yourself, then you may find the notes below helpful.

WHAT IS A LOCAL AREA NETWORK?
A local area network connects all your computers, printers etc. together.

WHY WOULD I WANT TO DO THAT?
Connecting your computers and printers together allows you to share your resources. For example; one printer can service many users – they don’t all need printers connected to their PC’s. Only one set of templates is required. Rather than every user having a number of templates on their PC, one central set of templates is shared between all users. Ensuring that everyone is using the same template guarantees a uniform look to your company’s documents. Sharing a single file is especially important with dynamic data such as in a database.

Consider this:-

Imagine that you have a simple database to store data about all your customers and orders. If two or more users need access to the database then they both need a copy of it on their PC’s. If one user changes something then the two copies will be different, until the files are merged and copied back to both users.

This problem disappears with a LAN, since both users can share the same files, also that data can only be accessed from those two particular PC’s which may be inconvenient. If you have more than a handful of people using PC’s in your company then installing a LAN can make user management and security easier to manage because you only need to set up user accounts for each user once on a central computer designated as a domain controller rather than having to set up access for each user on each PC, something which is usually overlooked;

PC’s are left open without any sort of access control making sensitive commercial data available to everyone in the company, not desirable in the light of a recent Computer Security Institute study which states that 70% of organisations polled had experienced data security incidents and that 60% of those were committed by internal staff!

WHAT DOES IT COST?
Local area networks are now so common that the cost is minimal. Somewhere in the order of £20 - £25 per PC, depending on the number of PC’s that you want to connect and how far apart they are.

WHAT DO I NEED?

Each device (PC, Printer etc.) requires a Network Interface Card (NIC). These are available at almost every computer store. They plug into one of the PC’s internal slots and are usually plug & play. You will need a switch or hub* into which they all plug and some patch cables (or a structured cabling infrastructure in a large multi office environment).

SWITCHES Vs HUBS:-
These are two devices for connecting computers together. A hub is rather like a main road with a number (usually 4, 8 or 16) of side roads connecting into it. There is no traffic control, therefore data joining the main highway from a side road often collides with data already on the highway. This data is damaged and lost. The device that sent the data will realise that this has happened because it won’t get a reply and will send the data again – maybe this time it will collide, maybe not. Following each collision the sending device waits longer before re-sending the data. As you can see this slows down the network, degrading performance significantly on a busy network. Switches are similar but have traffic control at each junction. Collisions should not occur. Additionally, hubs send all data to every device on the network. This means that each device is continually processing data destined for someone else. The devices have to inspect the IP headers and look at the destination address. If it is not for them they ignore it. This is a quick process, but it still has to be done and takes a finite time. Switches only send data to the device it is destined for, thus reliving the devices of unnecessary processing of network data. In summary switches make for a much more efficient network, and on busy networks considerably increase performance. They are, as you would expect more expensive than hubs.

LET NSI DO IT FOR ME!

HOW EASY IS IT?
Generally speaking, setting up a small network is reasonably easy. Getting it all to work usually involves fighting with windows to some extent, generally with the drivers for the NIC’s, but if you are used to installing devices in your home PC then you won’t have too much trouble. You should spend some time deciding who needs to access what, and take the trouble to set up security on the shares. Not difficult, but you will be surprised at how long you spend getting this bit right.

HOW TO DO IT

As stated above you need to install a NIC into each device to be connected. The devices are then all plugged into a switch or hub and the PC’s configured to use a common protocol, nowadays universally TCP/IP. Finally files and directories (Network Resources or resources for short) are shared by setting up sharing in windows explorer. Lets take a closer look at how to do this:

  • Install the NIC cards as per the manufacturers instructions
  • Open the network control applet in the control panel and select properties.
  • In the resulting dialogue box, tick “File & Print sharing”
  • In the TCP/IP section select the manual IP address option and enter a valid address and mask – you do not need to enter a gateway or DNS server setting.

NB Valid IP addresses refer to technically and legally valid. To ensure compatibility with Internet connections (Internet communication uses IP addresses) there are three ranges of private addresses. These are: -

  • 10.0.0.0 – 10.255.255.255
  • 172.16.0.0 – 172.31.255.255
  • 192.168.0.0 – 192.168.255.255

The subnet mask defines which part of the address is network and which is host, or computer. The subject of IP addresses and subnet masks gets rather technical, but I suggest that you use 255.255.255.0 as the mask on ALL your PC’s, and IP addresses of, say 10.0.0.1 – 10.0.0.2 etc. Each device MUST have a different address with the SAME mask and the first three octets (10.0.0) MUST be the same on each device.

IP addresses are numbers between 0 and 255. Using a mask of 255.255.255.0 means that you can’t use 0 or 255 in the last octet, any number in between is fine though.

In Windows explorer, right click the folder or file that you want to share and click on the ‘Sharing’ option. Enter a name for the share and that’s it!

NB: Windows NT, Windows 2000 and Windows XP all use a disk format called NTFS. This provides security features, which you can set so that only certain users can access the shared data. If you right click the newly shared folder and select properties and there is a security tab at the top then the PC uses NTFS and you can use this facility.

To access the share from another PC on the network open the ‘Network Neighbourhood’ icon on the desktop and you should see the computer names of the other devices on the network. Double clicking on these will show the resources shared by these devices. Double clicking the folders will open them and likewise the files within.

This is a brief overview of how to set up a basic network, sometimes it takes a bit of fiddling to get it all working and there is a lot of fine tuning that can be done in terms of security and functionality, see one of the many windows books for further details, and good luck.

WHAT ARE THE RISKS?
Generally there are few risks when setting up a small network. The obvious thing is that there are cables between every device that need to be well routed so as not to cause a trip hazard. Security of data is another concern, you may not want the new office assistant to be able to access confidential accounts data or buy prices etc. I strongly recommend that you use PC’s running one of the operating systems, which use NTFS so that you can control WHO has access to WHAT.

LET NSI DO IT FOR ME!

CONSIDERATIONS FOR LARGER NETWORKS:-
The network model described above is fine for a limited number of devices, say up to 50 or so PC’s. For larger networks it becomes necessary to use what is called a routed network. As mentioned earlier IP addresses have a network part and a host (device) part defined by the subnet mask. This is rather like a postal address where there is a house number (device part) and a street, town, county part (network part). Devices with the same network portion of an IP address are like houses in the same street.

When you connect to a device by name your PC needs to resolve that name to an IP address, further the IP address then needs to be resolved to a MAC address.

MAC addresses are the real addresses that devices use to communicate. They consist of six 2 byte hexadecimal numbers; something like “00-00-35-F4-C3-59”. Since it is impossible to ever remember these numbers for each device, we use logical addresses and names which we mere humans can handle more easily.

The way in which devices resolve these addresses is by sending a query to EVERY device on the network (network here refers to the number of hosts with the same network portion of the IP address). These queries are called broadcasts. There are a lot of these broadcasts going on in even a small network. If you were to set up a network with hundreds of devices there would be so many broadcasts traversing the network that there would be little time left for actual data!

Routed networks are networks where the IP addresses of devices span a number of different network portions. This means that each network portion has it’s own broadcasts, but these are not forwarded to devices on a different subnet. A subnet refers to a network which shares the same network portion of IP address. Also each subnet is said to be a separate broadcast domain, since broadcasts are contained within a subnet.

Using routed networks controls the broadcast problems and makes it easier to administer since they use layered addressing – like the house number and the street name idea mentioned earlier. It is easier to remember that all your account department PC’s are in a road with one name and your sales PC’s are in a different road. Immediately you can tell which function a device has just by the street name. Compare this to a flat addressing approach, like National Insurance numbers. Trying to remember 500 of those is much harder!

Routed networks use devices called routers to connect the various subnets. These devices look at the data and decide where to send it next so that it eventually arrives at the correct address. The ‘Default Gateway’ referred to in the PC’s IP set-up dialogue box is the IP address of that subnet’s router. Broadcast traffic will not traverse routers, this eliminates the problems of high broadcast traffic on a large network, but it means that the PC’s on different subnets cannot ‘see’ each other since they use broadcasts to resolve names! This problem is overcome by the use of either ‘Windows Internet Naming Service’ or WINS for short on Microsoft systems, or ‘Domain Name System’ or DNS for short on all other systems.

From the above it is clear that if your network grows to an extent where it becomes necessary to segment it using routers, you need to install a WINS or DNS server as well.

If there is some particular reason not to use WINS or DNS on your network – and I can’t think of one! - there is a file on every PC called ‘Hosts’. If all the IP addresses and device names of all the devices on the network are entered in this file then the PC will resolve names by looking them up in this file. Clearly each device on the network requires identical copies of this file for everything to work correctly, and if a change is made ALL the PC’s will need their files updated, quite a task!

It is usual practice to use a centralised configuration for all the PC’s on the network. Manually administering the settings of 50 or more PC’s can be quite a time consuming and ongoing process! There is a protocol called ‘Dynamic Host Configuration Protocol’ DHCP for short, which is used for this purpose. This basically dishes out IP address, Default Gateway, DNS server addresses and other configuration data to a PC each time the PC boots up. This is the process used when the ‘Obtain IP address and DNS server settings automatically’ option is ticked in the IP properties dialogue box of a PC’s network configuration.

In summary:

Use switches in preference to hubs – they considerably increase performance in a busy network.
If the network contains more than 40 –50 devices it is usual to use a routed network.
All but the smallest networks should have a Domain Controller server to centrally manage users and security.
Routed networks usually have a Name resolution server – WINS or DNS.
It is desirable to have a DHCP server in a network of this size or bigger.

LET NSI DO IT FOR ME!